// SCAN REPORT

Nodebench

824 tools. 312 can modify or destroy data without limits.

10 destructive tools with no built-in limits. Policy required.

Last updated:

312 can modify or destroy data
512 read-only
824 tools total

Community server · catalogue entry checked 28/06/2026

How to control Nodebench ↓

TOOL MAP

What Nodebench exposes to your agents

Read (512) Write / Execute (302) Destructive / Financial (10)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS
Critical Risk

The most dangerous Nodebench tools

Destructive · High deeptrace_revoke_passport Revoking an agent's passport is an irreversible action that permanently withdraws authorization or access credentials. Destructive · High delete_learning This tool irreversibly deletes data (a learning record) and cannot be undone. Destructive · High delete_sandbox_policy This tool permanently removes a sandbox policy configuration. Destructive · High deleteAgentMemory This tool performs irreversible deletion of agent memory entries.

312 of Nodebench's 824 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Nodebench

PolicyLayer is an MCP gateway — it sits between your AI agents and Nodebench, and nothing reaches the server without passing your rules. These are the rules we recommend:

Deny destructive operations
{
  "abandon_cycle": {
    "deny_if": [
      {
        "conditions": [],
        "on_deny": "Blocked by default. Requires approval."
      }
    ]
  }
}

Destructive tools should never be available to autonomous agents without human approval.

Rate limit write operations
{
  "accept_shared_task": {
    "limits": [
      {
        "counter": "accept_shared_task_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations
{
  "analyze_experiment_data": {
    "limits": [
      {
        "counter": "analyze_experiment_data_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

Controls API costs and prevents retry loops from exhausting upstream rate limits.

  1. Create a free account and register Nodebench — nothing to install.
  2. Add these rules — paste them, or build them visually. Tune the limits to your setup.
  3. Point your MCP client (Claude, Cursor, anything) at your gateway URL.
ENFORCE POLICY ON NODEBENCH →

Instant setup, no code required.

FULL CATALOGUE

All 824 Nodebench tools

DESTRUCTIVE 10 tools
Destructive abandon_cycle Abandon an active verification cycle that will not be completed. Use this to clean up orphaned or stale cycles Destructive archiveDocument Archive (soft-delete) a document and all its children recursively. Destructive cleanup_stale_runs Clean up orphaned eval runs stuck in Destructive deeptrace_revoke_passport Revoke an agent Destructive delete_learning Delete a learning by key. Use when a learning is outdated or incorrect. Destructive delete_sandbox_policy Delete a sandbox policy. Fails if any active sessions are using it. Destructive deleteAgentMemory Delete stored memory by key Destructive removeDocumentFromFolder Remove a document from a folder. Destructive share_revoke_packet_link Revoke a local share link so it no longer counts as active. Destructive unload_toolset Remove a dynamically loaded toolset from the current session to free up context. Cannot unload toolsets from t
EXECUTE 110 tools
Execute benchmark_models Run the same prompt against multiple LLM providers and compare responses. Returns side-by-side results with la Execute bootstrap_parallel_agents Detect whether a target project repo has parallel agent infrastructure and, if not, scaffold everything needed Execute build_banking_packet Build a banker-readiness packet from the canonical company packet. Execute build_company_profile_starter Build a starter PitchBook/Crunchbase-like company profile. Execute build_founder_operating_model Build the complete founder operating model: execution order, queue topology, packet routing, source trust poli Execute build_research_digest Generate a digest of new (unseen) articles from RSS feeds. Compares against previously seen articles via SQLit Execute build_shared_context_subscription Build the exact pull/subscription manifest an agent client should use to watch a packet or packet scope. Execute build_shared_context_subscription_manifest Build a filtered snapshot/events/pull manifest for one peer, packet class, producer, scope, or subject so clie Execute build_submission_export Build a generic submission export from the canonical company packet. Execute build_temporal_graph Build a temporal relationship graph for an entity. Execute burst_capture Capture N sequential screenshots at fixed intervals using Playwright. Execute call_driver_tool Invoke a tool on a connected MCP driver. This proxies the call to the external MCP server (e.g. playwright-mcp Execute call_llm Call an LLM model directly and get the response with metrics (tokens, latency). Uses available API keys: Gemin Execute call_openclaw_skill Run an OpenClaw skill safely through security checks. Execute call_webmcp_tool Invoke a WebMCP tool on a connected origin. The tool is executed in the browser page context via page.evaluate Execute capture_surface_stats Capture Android SurfaceFlinger stats and logcat for jank analysis (Layer 0 only). Returns janky frame counts, Execute capture_ui_screenshot Capture a screenshot of a URL using headless Playwright. Returns the screenshot as an inline image that multim Execute compile_decision_packet Compile entity intelligence into a decision-ready packet. Execute compile_scenarios Generate 3-7 future scenario branches for an entity or decision. Execute compile_tension_model Model explicit tensions between forces for a decision or entity. Execute compute_ssim_analysis Compute block-based SSIM analysis on a set of frame images. Uses 8x8 blocks with parallel ProcessPoolExecutor. Execute connect_mcp_driver Connect to an external MCP server and make its tools available through nodebench-mcp. Predefined drivers: Execute connect_openclaw Connect to an OpenClaw agent with a security policy applied. Execute connect_webmcp_origin Connect to a WebMCP-enabled website via Playwright. Navigates to the URL, intercepts navigator.modelContext to Execute convex_pre_deploy_gate Run a comprehensive pre-deployment quality gate. Checks: convex/ directory structure, schema.ts validity, depr Execute convex_quality_gate Run a configurable quality gate across all stored audit results. Like SonarQube Execute delta_self_dogfood Dogfood NodeBench Delta on itself. Verifies runtime health, setup friction, distribution surfaces, and compoun Execute disconnect_driver Disconnect from an external MCP driver and shut down its child process. Use this to clean up or to reconnect w Execute disconnect_openclaw Disconnect an OpenClaw session and generate a safety summary. Execute disconnect_webmcp_origin Disconnect from a WebMCP origin and close the browser page. Use this to clean up resources or to reconnect wit Execute dive_auto_discover Scan the current page DOM and auto-register components in the dive tree. Discovers semantic landmarks (nav, he Execute dive_fix_verify After fixing a bug, verify the fix by re-navigating to the affected route, comparing before/after state, and u Execute dive_interaction_test Define and track a structured interaction test for a component. Provide preconditions and a sequence of test s Execute dive_reexplore Re-traverse a route after code changes to detect regressions and verify fixes. Compares the current state agai Execute end_openclaw_session End an OpenClaw session and generate a safety summary. Execute enforce_merge_gate Pre-merge validation combining git state, verification cycles, eval runs, test results, and quality gates. Ret Execute extract_video_frames Record screen and extract key frames from an Android device (Layers 1+2). Uses adb screenrecord then ffmpeg sc Execute founder_local_weekly_reset One-call convenience: gathers all local context and produces a complete Execute get_path_replay Replay a session Execute gtm_script_builder Build a starter GTM script for the current founder wedge. Execute invoke_openclaw_skill Run an OpenClaw tool safely through security checks. Execute invoke_view_tool Invoke a per-view tool on the current or specified view. Execute judge_request_retry Request a retry, re-plan, escalation, or stop for a failed subtask. Execute judge_tool_output Run the 7-criterion LLM judge on a tool Execute load_toolset Dynamically load a toolset into the current session. After loading, the tools become immediately available for Execute log_interaction Log and optionally auto-execute an interaction step. If the built-in Playwright browser is active (launched by Execute manipulate_screenshot Manipulate a screenshot using sharp (image processing). Supports crop (extract a region), resize, and annotate Execute navigate_to_view Navigate to a specific view in the NodeBench AI frontend. Execute nb_start_agent Start a new agent conversation with an optional initial message. Execute nb_switch_research_tab Switch the research hub tab. Execute nodebench_ask_agent Send a question to the NodeBench AI agent. Returns a structured response with reasoning and sources. Execute nodebench_navigate Navigate to a specific view. Returns the target view Execute nodebench.claims.verify Re-run deterministic public/private boundary and source-evidence verification for a stored public claim. Execute nodebench.research_role Run or reuse public role research, store public hiring or market claims, and return a compact role dossier. Execute nodebench.research_run Run adaptive, evidence-backed research across one or more subjects. Automatically resolves entities, infers sc Execute open_dive_dashboard Open the NodeBench UI Dive dashboard in a browser. Shows the full flywheel cycle: Execute open_local_dashboard Start the local Daily Brief dashboard server if needed, and return the URL. The dashboard shows Brief metrics, Execute project_financials Build 5-year financial projections based on historical data and industry assumptions. Projects: - Revenue wit Execute readiness_scan Run a founder readiness scan against the progression and diligence model. Execute render_flow_visualization Render flow visualization with colored bounding boxes for each flow group. Supports overlay on a rendered page Execute request_execution_approval Request a human approval gate for a risky execution-trace action. Approval state is written onto the live run Execute run_autonomous_loop Execute autonomous verification loop with stop conditions. Implements Ralph Wiggum pattern with checkpoints, i Execute run_benchmark_batch Run a longitudinal benchmark batch. N=1 is a smoke test (1 founder, 1 session). Execute run_browserstack_benchmark_lane Return a BrowserStack/browser-automation benchmark lane payload. Execute run_closed_loop Track a compile-lint-test-debug closed loop iteration. Record the result of each step. Never present changes w Execute run_code_analysis Static analysis on code or text content for security issues, secrets, homograph attacks, ANSI injections, susp Execute run_competitor_signal_benchmark Return a competitor-signal-to-response benchmark lane payload. Execute run_deep_sim Run a multi-agent scenario simulation with bounded branching and budget controls. Instantiates agents with per Execute run_dogfood_batch_with_judge Execute the priority 3 dogfood scenarios with automatic LLM judge validation. Execute run_entity_intelligence_mission Run a full DeepTrace entity intelligence mission with optional bounded research cell. Unifies relationship map Execute run_flicker_detection Run full 4-layer Android UI flicker detection pipeline: SurfaceFlinger stats + logcat (L0), screenrecord (L1), Execute run_founder_autonomy_benchmark Run the weekly founder reset autonomy benchmark lane. Execute run_graphify Generate a knowledge graph from a folder of code, docs, papers, or images. Execute run_judge_loop Execute a full judge-fix-verify loop: calls a tool, judges the output, and if it fails, Execute run_mandatory_flywheel Enforce the mandatory 6-step AI Flywheel verification after any non-trivial change. All 6 steps must pass befo Execute run_oracle_comparison Compare actual output against a known-good oracle reference. Based on Anthropic\ Execute run_packet_to_implementation_benchmark Return a packet-to-implementation benchmark lane payload. Execute run_quality_gate Evaluate content or code against a set of boolean rules. Returns pass/fail with specific failures listed. The Execute run_recon Start a reconnaissance research session. Use this at the start of Phase 1 (Context Gathering) to organize rese Execute run_research_cell Run a bounded re-analysis cell for a DeepTrace entity investigation. Queries existing DeepTrace state through Execute run_self_directed_delivery_loop Run a local-first autonomous delivery loop across exploratory research, planning, implementation commands, dog Execute run_self_heal Autonomous self-healing for detected drift issues. Fixes orphaned verification cycles Execute run_self_maintenance Run autonomous self-maintenance cycle. Checks TypeScript compilation, documentation sync, tool counts, test co Execute run_signal_sweep Run a live signal sweep across all data sources (HackerNews, GitHub Trending, Yahoo Finance, ProductHunt). Ret Execute run_sync_bridge_flush Open the outbound websocket bridge, pair or resume the local device, and flush pending approved operations to Execute run_tests_cli Execute a shell test command with timeout, capture stdout/stderr, and return structured results. Useful for ru Execute run_visual_qa_suite End-to-end visual QA pipeline: burst capture → SSIM stability analysis → Execute runNewsroomPipeline Trigger the full DRANE newsroom pipeline (Scout > Historian > Analyst > Publisher) for a given topic or entity Execute runSpotFixScan Scan for common operational issues: stale missions, blocked tasks with met deps, old sniff checks. Execute sandbox_batch Execute multiple commands, index all outputs, and run multiple search queries — all in ONE call. This is the h Execute sandbox_execute Run a shell command, automatically index the output into the sandbox, and return only a summary. The raw stdou Execute scrapling_crawl Start a multi-page spider crawl with extraction. Crawls from start URLs, follows links matching a CSS selector Execute scrapling_crawl_stop Stop a running crawl session. Pass the session_id from scrapling_crawl. Items collected so far are preserved. Execute self_implement Self-implement missing agent infrastructure. Generates implementation plan and code templates for: agent_loop, Execute simulate_decision_paths Run Monte Carlo simulation for founder decisions. Generates multiple random paths to visualize possible future Execute smart_select_tools LLM-powered tool selection: sends your task description + a compact tool catalog to a fast model (Gemini 3 Fla Execute spawn_openclaw_agent Start a secure OpenClaw session with safety rules applied. Execute start_autonomy_benchmark Start an autonomous capability benchmark. Defines a complex build challenge and tracks agent progress through Execute start_dogfood_session Start a new dogfood session for one of the 3 canonical loops (weekly_reset, pre_delegation, company_search). R Execute start_eval_run Start a new eval run. Define the test batch upfront with test cases (input, intent, expected behavior), then r Execute start_execution_run Start a live Convex-backed execution trace run for a workflow. Creates a task session and trace together so la Execute start_ui_dive Initialize a UI/UX Full Dive session. Auto-launches a headless Playwright browser if installed (zero setup). N Execute start_verification_cycle Start a new 6-phase verification cycle for a non-trivial implementation. Returns the cycle ID and Phase 1 inst Execute switchTab Switch between research hub tabs Execute thompson_pipeline End-to-end Thompson Protocol pipeline orchestrator. Takes a complex topic and runs it through all 4 agents (Wr Execute transcribe_audio_file Transcribe a local audio file (MP3/WAV/etc) to text using faster-whisper via Python. Deterministic, no network Execute trigger_batch_run Run a scheduled task right now instead of waiting. Execute trigger_investigation When an eval run shows regression, trigger a new verification cycle to investigate. This is how the outer loop Execute triple_verify Run triple verification on agent implementation. V1: Internal codebase analysis. V2: External authoritative so Execute validate_agent_compatibility Run the agent validation harness — simulates how AI agents (Claude Code,
WRITE 192 tools
Write accept_shared_task Accept a proposed shared-context task. Write ack_shared_context Acknowledge that a peer received and accepted a context packet. Write add_forecast_evidence Add evidence to a forecast Write add_rss_source Register an RSS or Atom feed URL for monitoring. Stored in SQLite for persistent tracking. Validates the feed Write addDocumentToFolder Add a document to a folder. Write archive_content Save generated content to the archive for deduplication and theme tracking. Prevents the engine from regenerat Write assign_agent_role Assign a specialized role to the current agent session. Roles define focus area and behavioral instructions. P Write attach_execution_evidence Attach evidence to a live execution trace. Use for URLs, uploaded files, screenshots, render outputs, and trut Write bind_local_account Record explicit local pairing permission so this device can map durable local context to a specific web user a Write bootstrap_project Register or update your project Write broadcast_agent_update Broadcast a status update to all active agents. Unlike send_agent_message (point-to-point), this creates a mes Write build_before_after_memo Build a memo showing the before and after path plus the validation rationale. Write build_causal_chain Construct a causal chain from temporal observations. Nodes must be in chronological order. Each node represent Write build_company_packet Build the canonical company readiness packet. Write build_slack_onepager Build a Slack-friendly one-page founder report. Write claim_agent_task Claim a task lock so other parallel agents know you Write claimTask Claim a pending task for execution. Returns the task if successful, null if already claimed or dependencies no Write compile_environment_spec Generate a simulation environment specification from entity intelligence. Write complete_autonomy_benchmark Finalize an autonomy benchmark run. Computes final score, duration, tool usage stats, and comparison against r Write complete_eval_run Finalize an eval run and compute aggregate scores. Returns pass rate, average score, failure patterns, and imp Write complete_execution_run Finish a live execution run by updating session status, trace status, and optional usage metrics. Use this aft Write complete_shared_task Complete a shared-context task and attach the output packet if one was produced. Write compress_or_expand_text Precisely compress or expand academic text by a target word count. Compress mode: remove filler words, convert Write compute_dimension_profile Recompute and persist the DeepTrace dimension profile for an entity. Use after new company evidence, relations Write configure_channel_preferences Set your messaging preferences: which channels to use first, Write configure_sandbox_policy Create or update a security policy for OpenClaw agent sessions. Write convex_audit_transaction_safety Audit Convex functions for transaction safety: multiple ctx.runMutation calls in actions (separate transaction Write convex_generate_plan Generate a Convex-specific implementation plan for missing code signatures. Takes the gap analysis from convex Write convex_generate_rules_md Generate or update a Convex rules markdown file from the current gotcha database, recent audit results, and pr Write convex_record_gotcha Record a Convex-specific gotcha, edge case, or pattern for future reference. Stored persistently and searchabl Write convex_schema_migration_plan Compare two schema snapshots and generate a migration checklist. Shows added/removed tables, index changes, an Write create_forecast Create a new forecast with a question, resolution date, and criteria. Optionally set initial probability, base Write create_proof_pack Assemble an immutable proof pack for verification. Bundles a checklist (pass/fail items), optional metrics (to Write create_task_bank Create or add to a fixed task bank for controlled agent evaluation. Each task defines: initial state (repo sna Write create_visual_pr End-to-end PR creation: exports screenshots, generates a rich markdown PR body with visual evidence (before/af Write createDocument Create a new rich-text document with a title and optional initial content blocks. Write createDocumentWithContent Create a document with pre-built ProseMirror JSON content string. Write createFolder Create a new folder for organizing documents. Write createMission Create a new mission with title, type, success criteria, output contract, and budget. Write createPlan Create an explicit task plan as a markdown document with steps marked as pending/in_progress/completed Write createSpreadsheet Create a new spreadsheet with a name. Write decide_re_update Decide whether to update existing instructions or create new files. Implements Write deeptrace_create_evidence_pack Bundle multiple evidence chunks into a named evidence pack. Evidence packs are the unit of provenance — they c Write deeptrace_create_passport Create or update an agent passport — the scoped identity that defines what an agent can read, spend, sign, and Write deeptrace_ingest_evidence Ingest a piece of evidence (article, filing, data point) into the evidence store. Returns a content-addressed Write deeptrace_log_receipt Log a tamper-evident action receipt. Every agent action should produce a receipt recording what was done, what Write delegate_founder_issue Create a bounded shared task handoff for a founder issue packet so the weak angle becomes assigned work. Write delta_handoff Generate a delegation packet for handing off work to another agent or teammate. Produces a delta.handoff packe Write delta_memo Create a decision-ready memo artifact. Produces a delta.memo packet with recommendation, variables, scenarios, Write delta_retain Preserve context for future sessions. Produces a delta.retain packet storing important notes, decisions, meeti Write dismiss_alert Dismiss an important change alert so it no longer appears in proactive alerts. Sets the status to Write dive_changelog Record a change made to fix a bug, design issue, or improve a component. Links before/after screenshots to sho Write dive_design_issue Tag a design inconsistency found during the dive. Covers visual problems like color mismatches, spacing deviat Write dive_generate_tests Generate Playwright regression test code from dive findings. Creates test cases from: bugs (verify the fix hol Write dive_link_backend Link a UI component to its backend dependencies. Connect components to API endpoints, Convex queries/mutations Write dive_record_test_step Record the actual result of a test step after executing it via the MCP Bridge. Compare expected vs actual, att Write dive_save_screenshot Save a screenshot during a dive session. Pass base64 image data (from bridge Write draft_email_reply Structure an email thread for reply drafting. Parses the thread, extracts context (from, subject, date), and b Write duplicateDocument Duplicate a document. Clones content, icon, type. Resets visibility and favorite. Write end_component_flow Complete a component Write end_dogfood_session End a dogfood session with summary metrics: time-to-first-useful-output, delegation success, packet export sta Write escalate_shared_task Escalate a shared-context task when the assignee cannot complete it cleanly. Write export_artifact_packet Formats a Founder Artifact Packet or memo for export to a specific audience and format. Write flag_important_change Flag a detected important change with impact scoring, affected entities, and optional suggested action. Used b Write founder_local_synthesize Takes gathered local context and synthesizes a complete Founder Artifact Packet. Write generate_academic_caption Generate academic figure or table captions following top-venue conventions. Handles Title Case for noun phrase Write generate_flicker_report Generate visual flicker report from existing analysis data. Produces SSIM timeline chart (1200x400 PNG, PIL-ba Write generate_grid_collage Tile N screenshot images into a single grid collage PNG for visual inspection. Write generate_implementation_plan Generate a structured implementation plan for missing code signatures. Takes the gap analysis from verify_conc Write generate_parallel_agents_md Generate a portable, framework-agnostic AGENTS.md section for parallel agent coordination. Designed to be drop Write generate_plan_delegation_packet Convert a FeaturePlan into an agent-ready delegation packet Write generate_pr_report Generate a rich markdown PR body from a UI Dive session. Compiles visual changes (before/after screenshot comp Write generate_proposal_memo Render a FeaturePlan as a human-readable proposal memo. Write generate_report Compile structured findings, eval results, and quality gate data into a formatted markdown report. Useful for Write generate_self_instructions Generate self-instructions for the agent in various formats: skills_md (SKILL.md), rules_md (RULES.md), guidel Write generate_team_install_plan Generate a practical install and rollout plan for a founder, solo developer, or small team using NodeBench MCP Write generate_voice_scaffold Generate starter code for a voice bridge. Returns file contents, setup instructions, and dependency lists for Write generate_zero_draft Auto-draft an artifact (slack message, email, spec doc, PR draft, architecture note, career plan, or content b Write graphify_import_to_subconscious Import a graphify knowledge graph into NodeBench Write ingest_dive_screenshots Scan a directory for PNG/JPG screenshot files and bulk-import them into the dive session Write ingest_temporal_observation Ingest a raw observation into the temporal substrate (timeSeriesObservations). Supports numeric, categorical, Write ingest_upload Ingest uploaded file content into the NodeBench entity intelligence system. Write install_nodebench_plugin Generate or write a starter .mcp.json entry for NodeBench MCP so a local team member can install the preset qu Write judge_session Score a dogfood session on 6 dimensions (1-5 each): truth, compression, anticipation, output, delegation, trus Write link_durable_objects Create a durable relationship such as screen -> action, workflow -> run, run -> artifact, or outcome -> eviden Write log_benchmark_milestone Record completion of a benchmark milestone. Tracks which milestones the agent achieved, time taken, tools used Write log_gap Record a gap found during Phase 2 (Gap Analysis). Gaps are categorized by severity: CRITICAL (protocol violati Write log_phase_findings Record findings for the current phase of a verification cycle. Advances the cycle to the next phase if the cur Write log_recon_finding Record a finding from reconnaissance research. Link it to a recon session and categorize it. Use for both exte Write log_test_result Record a test result for Phase 4 (Testing & Validation). Tests are organized by layer: static, unit, integrati Write manage_implementation_packets Create and manage implementation packets — structured instructions for Claude Code or other coding agents. Write manage_task_list Manage the workspace task list. Add, update, complete, delete, or list tasks. Write merge_compose_output Judge-gated merge of subtask artifacts into a composed output. Write merge_research_results Merge parallel sub-agent research results into a unified dataset. Takes arrays of records from multiple source Write nb_open_approval_queue Open the approval queue for held actions. Write nodebench_create_document Create a new document in NodeBench. Supports markdown content with optional metadata tags. Write nodebench.capture Persist a messy event capture into the active NodeBench event workspace without live paid search. Uses event c Write nodebench.claims.submit_public Submit a sourced public claim. The verifier rejects private sources, raw email/resume/private artifact text, a Write nodebench.link_private_signal_to_public_entity Link an app-private signal to a public entity by hash only. Raw Gmail, resume, inbox, and private artifact tex Write nodebench.notebook_append Append reviewed text into a NodeBench report notebook through the same Convex-backed report notebook persisten Write nodebench.report_export_complete Complete a previously previewed NodeBench report export after review. Writes the export completion event to th Write nodebench.submit_public_claim Submit a sourced public claim to NodeBench. The verifier rejects private email/resume-derived text and non-pub Write nodebench.watch_entity Create a watch request for an entity. Current MVP returns the entity and recommended refresh cadence. Write open_core_boundary_advisor Advise what should stay open-core versus proprietary. Write plan_decompose_mission Decompose a mission into subtasks with verifiability routing. Write polish_academic_text Deep-polish academic text for top-venue quality (NeurIPS, ICLR, ICML, ACL). Handles English and Chinese papers Write promote_to_eval Take findings from a completed verification cycle and promote them into eval test cases. This is how the inner Write propose_shared_task Propose a task handoff between peers with input contexts and required output packet shape. Write publish_founder_issue_packet Turn the weakest founder-direction angle into a durable shared-context issue packet with lineage, proof links, Write publish_shared_context Publish a structured shared-context packet with subject, claims, evidence refs, freshness, permissions, and li Write publish_to_queue Push content to the LinkedIn content queue on the Convex platform. Content goes through the engagement gate an Write queue_sync_operation Queue an explicit outbound sync operation when a custom workflow needs to push metadata, receipts, or approval Write rate_packet_usefulness Rate a packet Write record_dogfood_telemetry Record a full telemetry row for a dogfood run. Captures surface, scenario, user role, prompt, tool usage, toke Write record_eval_result Record the actual result for a specific eval case. Include what happened, the verdict (pass/fail/partial), and Write record_event Record a typed event to the causal event ledger. Supports causal linking via causedByEventId and correlation g Write record_execution_decision Record a structured decision on a live execution trace without storing raw hidden reasoning. Use for rankings, Write record_execution_step Record a structured execution step receipt on a live execution trace. Use this for meaningful actions like fil Write record_execution_verification Record a verification result on a live execution trace. Use for render checks, formula checks, diff checks, ar Write record_fix_attempt Record a fix attempt with replay proof and regression protection description. Links to a failure case. Write record_learning Store an edge case, gotcha, pattern, or regression discovered during verification. Learnings are searchable vi Write record_manual_correction Track a human correction to agent output. Every correction is evidence of a system gap — the system should hav Write record_openclaw_gotcha Record a discovered OpenClaw pitfall or security finding. Write record_path_step Record a navigation/exploration step in the user Write record_provenance_receipt Persist a durable execution receipt for a tool call, approval, verification, or other meaningful action. Write record_repeated_question Track a question the user asked that NodeBench should have already known. This is the core failure signal — re Write record_state_diff Record a before/after state change on an entity. Tracks what changed, which fields, and why. Write record_sync_artifact Persist a local artifact with verification state so it can be replayed, reviewed, and optionally synced to the Write record_sync_outcome Persist an outcome with user value, stakeholder value, evidence, and status so the system always resolves work Write recordCanaryRun Record a new canary benchmark run with throughput and quality scores. Write register_component Register a UI component in the dive tree. Components form a hierarchy: page → section → form/modal/list → butt Write register_shared_context_peer Register a scoped peer with product, role, surface, capabilities, and heartbeat metadata for shared-context co Write register_skill Register a skill (rule/memory .md file) with its source documents, update triggers, Write reject_shared_task Reject a proposed shared-context task with a reason. Write release_agent_task Release a task lock after completing work. Updates status and optionally records a progress note for the next Write remove_ai_signatures Detect and remove AI-generated writing signatures from academic text. First runs pattern matching for known AI Write report Produce a human-readable artifact for either a research topic or a decision. If recommendation inputs are prov Write resolve_forecast Resolve a forecast with an outcome. Auto-computes Brier and log scores for binary forecasts. Ambiguous outcome Write resolve_founder_issue Invalidate a founder issue packet and optionally publish a resolution packet so the issue lifecycle stays expl Write resolve_gap Mark a gap as resolved after implementing the fix. Returns remaining gap counts by severity. Write resolveRoutingRecommendation Accept, reject, or expire a routing recommendation. Write resolveSniffCheck Resolve a pending sniff check with approved, rejected, or needs_revision. Write restoreDocument Restore an archived document and its children. Write retention_register_connection Register a retention.sh team connection in local MCP state so QA findings and token savings can flow into foun Write retention_sync Sync data between NodeBench Delta and retention.sh. Pushes delta packets as team context and pulls QA findings Write retention_sync_findings Sync retention.sh QA findings, scores, and token savings into local MCP state. Write sandbox_ingest Index arbitrary text into the context sandbox (FTS5). Raw content stays in SQLite — only a compact reference e Write save_research_resource Save a research resource with URL, source citation, tags, and notes. Write save_session_note Persist a critical finding, decision, or progress note to the filesystem. Notes survive context compaction — c Write scaffold_directory Scaffold directory structure following OpenClaw patterns. Creates organized subdirectories and placeholder fil Write scaffold_nodebench_project Create a complete project template pre-configured for nodebench-mcp. Generates: package.json, AGENTS.md, .mcp. Write scaffold_openclaw_project Generate a starter project for OpenClaw + NodeBench. Write scaffold_openclaw_sandbox Generate Docker/WSL2/Podman setup files for running OpenClaw in an isolated environment. Write scaffold_research_pipeline Generate a complete, standalone Node.js project for an automated research digest pipeline. Creates: package.js Write send_agent_message Send a message to another agent by session ID or role. Enables asynchronous inter-agent communication for task Write send_email Send an email via SMTP over TLS. Requires EMAIL_USER and EMAIL_PASS env vars. Defaults to Gmail SMTP (smtp.gma Write send_openclaw_message Send a message through any connected channel. Write send_peer_message Send a direct structured message to a peer without routing everything through a central orchestrator. Write sendMessage Send a message to the live chat run Write set_watchdog_config Configure the background watchdog that continuously monitors system health. Write setup_operator_profile Set up your profile to customize how the AI assistant works for you. Write share_create_packet_link Create a durable local share link record for a packet or founder memo so it can be rendered or synced later. Write sniff_record_human_review Record a human sniff-check for a subtask or merge output. Write start_component_flow Claim a component for traversal by a specific subagent. Marks it as Write summarize Turn raw context into a compact brief with key points and optional persistence. This is the fast human-readabl Write sync_company Push a company profile into NodeBench AI from Claude Code. Extracts company truth from a summary you provide ( Write sync_daily_brief Sync daily brief + narrative data from Convex to local SQLite. Requires CONVEX_SITE_URL and MCP_SECRET environ Write sync_operator_profile Sync the Operator Profile to the local filesystem at ~/.nodebench/USER.md. Write sync_report Push a report artifact from Claude Code into NodeBench AI. The report is saved locally and published as a shar Write sync_skill Resync a stale skill after applying updates. Recomputes source hashes, updates Write synthesize_extension_plan Synthesize a plan for extending or deepening an existing feature. Write synthesize_feature_plan Synthesize a phased feature implementation plan conditioned on founder context, Write synthesize_integration_proposal Synthesize an integration plan for an external tool, API, or framework. Write synthesize_recon_to_learnings Convert recon findings into persistent learnings. Recon findings are ephemeral research notes; learnings are t Write tag_ui_bug Tag a bug to a specific component (and optionally a specific interaction). Bugs are categorized by severity (c Write thompson_write Transform complex content into Thompson Protocol format — plain English mandate, intuition-before-mechanics, a Write track Add, check, remove, or list tracked entities with one workflow tool. The default path optimizes for watched en Write track_action Record any significant action with before/after state, reasoning, and temporal metadata. Auto-captures session Write track_intent Track a user intent that should survive context window compaction. On Write track_milestone Record a significant milestone (phase complete, deploy, ship, launch, pivot, decision) with optional evidence Write update_agents_md Read, append, or update sections in the AGENTS.md file. This file contains instructions for AI agents working Write update_company_truth Update a subconscious memory block with new information. Write update_forecast_probability Update a forecast Write updateDocument Update document fields: title, content, icon, visibility, favorite status. Write updatePlanStep Update the status or notes of a specific step in a task plan Write upsert_durable_object Register or update a durable local object so views, tools, workflows, runs, artifacts, and outcomes share one Write upsertEntity Create or update a canonical entity record (company, person, fund, etc.). Write watchlist_add_entity Add an entity to the local founder watchlist with alert preferences and optional strategic-angle linkage. Write watchlist_refresh_entities Refresh watchlist timestamps and optionally attach change summaries for watched entities. Write write_workspace_file Create or update a file in the agent workspace (~/.nodebench/workspace/). Write writeAgentMemory Store intermediate results or data for later retrieval. Use this to avoid context window overflow. Write zip_extract_file Extract a single file from a local ZIP archive to a local output directory (zip-slip safe). Deterministic, no
READ 512 tools
Read analyze_experiment_data Analyze experiment data (CSV or JSON) and generate publication-ready analysis paragraphs. Uses \\paragraph{Con Read analyze_figma_flows Full Figma flow analysis pipeline: extract frames from Figma file (depth=3 tree traversal), cluster into flow Read analyze_repo Analyze a GitHub repository Read analyze_screenshot Send a screenshot to a vision-capable AI model for analysis. Accepts base64 image data (from capture_ui_screen Read analyze_seo_content Analyze HTML or text content for SEO quality: word count, sentence count, paragraph count, Flesch-Kincaid read Read analyze_voice_config Validate a voice pipeline configuration. Checks STT/TTS/LLM compatibility, estimates per-minute costs, identif Read ask_context Ask against saved NodeBench context first. This searches session memory and the accumulated knowledge base so Read assess_risk Assess risk tier for a given action. Returns tier (low/medium/high), reversibility, external impact, and recom Read audit_openclaw_skills Scan installed OpenClaw tools for security risks: dangerous patterns, Read audit_workflow_definition Analyze an OpenClaw workflow configuration for reliability, security, and performance issues: Read auditTelemetry Audit inference call telemetry for anomalies: cost spikes, latency outliers, zero-cost suspicious calls. Read benchmark_voice_latency Calculate theoretical latency for one or more voice pipeline configurations. Breaks down STT, LLM (first token Read build_claim_graph Extract claims from a source packet and link each claim to its evidence. Returns a directed graph of claims wi Read build_diligence_packet Build a diligence-oriented export payload from the canonical company packet. Read build_investor_packet Build an investor-oriented export payload from the canonical company packet. Read burn_rate_sanity Sanity check founder burn against runway and stage expectations. Read calculate_terminal_value Calculate terminal value using perpetuity growth or exit multiple method. Perpetuity method: TV = FCF(n+1) / Read calculate_valuation Calculate final enterprise value and equity value. Calculation: 1. PV of forecast FCF (Years 1-5) 2. PV of te Read calculate_wacc Calculate Weighted Average Cost of Capital (WACC) using CAPM. Components: - Risk-free rate (US 10Y Treasury) Read capture_responsive_suite Capture screenshots at all 3 standard responsive breakpoints (mobile 375px, tablet 768px, desktop 1280px) in o Read cheapest_valid_path Return the cheapest valid workflow path for the stated objective. Read check_agent_inbox Read unread messages for the current agent session. Filter by category, sender, or priority. Messages are mark Read check_contract_compliance Analyze an agent session Read check_design_compliance Check a single .tsx/.ts file for design governance compliance. Runs all banned pattern checks (color, typograp Read check_dive_drivers Setup wizard for MCP automation drivers. Probes the system for Android SDK (ANDROID_HOME, adb, emulator), iOS Read check_email_setup Diagnostic wizard for email tool configuration. Checks env vars (EMAIL_USER, EMAIL_PASS, etc.), optionally tes Read check_framework_updates Get a structured checklist of sources to check for framework/SDK updates. Pre-built source lists for: anthropi Read check_git_compliance Validate branch state, uncommitted changes, and conventional commit compliance. Checks if on a protected branc Read check_mcp_setup Comprehensive diagnostic wizard for the entire NodeBench MCP. Checks all env vars, API keys, optional npm pack Read check_openclaw_setup Check if OpenClaw is ready to use: verifies the server is installed, Read check_openclaw_status Check if OpenClaw is installed, Docker is available, and any sessions are active. Read check_page_performance Lightweight page performance check via HTTP fetch (no browser). Measures: response time, content size, compres Read check_paper_logic Check academic text for logical issues: contradictions between statements, undefined terms, terminology incons Read check_peer_messages Read the direct messages for a peer, optionally restricting to unread items. Read check_plugin_update_readiness Check whether NodeBench MCP is ready for a version/update push across installer, plugin metadata, and editor s Read check_skill_freshness Check if registered skills are stale by comparing current source file hashes Read check_webmcp_setup Check WebMCP prerequisites: Playwright installation, Chromium browser availability, and any cached origin data Read check_wordpress_site Detect whether a site runs WordPress and assess its security posture. Checks: WP generator meta tag, wp-conten Read claim_verification_scan Scan high-risk claims against available evidence classes. Read classify_failure Classify a failure by canonical system layer taxonomy. Tracks symptom, root cause, and system layer for struct Read cluster_figma_flows Cluster Figma frames into flow groups using multi-signal priority cascade: 1) section-based grouping (highest Read compare Compare 2-4 entities using the same report-shaped output. The tool backfills quick diligence packets first so Read compare_eval_runs Compare two eval runs to decide whether a change should ship. Returns side-by-side scores and a deploy/revert Read compare_founder_directions Compare multiple founder directions side by side across strategic angles, issue count, confidence, and recomme Read compare_options Generate a formatted side-by-side comparison table from scored research results. Takes either raw options or a Read compare_savings Compare token usage, time savings, and cost estimates. Optionally compare two sessions for before/after ROI. S Read compare_workflow_paths Compare current and optimized workflow paths and quantify likely savings. Read compute_calibration Compute calibration bins and Brier aggregates across all resolved forecasts. Returns 10 bins (0-10%, 10-20%, . Read compute_web_stability Compute frame-to-frame stability metrics for a burst capture. Calculates Read connect_channels Connect to multiple information channels for aggressive information gathering. Channels: slack, telegram, disc Read convex_analyze_components Analyze Convex component imports in convex.config.ts. Lists all installed Convex components (agent, workflow, Read convex_analyze_http Analyze convex/http.ts for HTTP endpoint issues: duplicate routes, missing CORS headers, missing OPTIONS prefl Read convex_audit_actions Audit Convex actions for: ctx.db access (fatal — actions cannot access DB directly), missing Read convex_audit_authorization Audit Convex functions for authorization issues: public mutations/actions writing to DB without auth checks, g Read convex_audit_data_modeling Audit Convex schema for data modeling issues: deeply nested objects (flatten into tables), dangling v.id() ref Read convex_audit_dev_setup Audit Convex project development setup: .gitignore includes _generated/, .env.example exists, convex.json is v Read convex_audit_diff Compare the latest audit run against the previous run to show new issues, fixed issues, and trend direction (i Read convex_audit_functions Audit all exported Convex functions for: missing validators (args/returns), old syntax usage, sensitive functi Read convex_audit_pagination Audit Convex pagination patterns: missing paginationOptsValidator in query args, unbounded numItems (DoS risk) Read convex_audit_query_efficiency Audit Convex queries for efficiency issues: .collect() without limits (unbounded reads), .filter() without ind Read convex_audit_schedulers Audit Convex scheduled function usage (ctx.scheduler.runAfter/runAt): detects infinite self-scheduling loops, Read convex_audit_schema Analyze a Convex project Read convex_audit_storage_usage Audit Convex file storage usage: missing null checks on ctx.storage.get()/getUrl(), storage IDs stored as v.st Read convex_audit_vector_search Audit Convex vector search implementation: validates vectorIndex dimensions against known embedding models, ch Read convex_bootstrap_project Scan a Convex project directory and generate a comprehensive health report. Checks: schema presence, function Read convex_check_crons Validate cron job definitions in convex/crons.ts. Checks: handler references use internal (not api), no duplic Read convex_check_env_vars Check that all environment variables referenced in Convex code (process.env.) are defined in .env files. Ident Read convex_check_function_refs Validate function references across the Convex codebase. Checks that ctx.runQuery/runMutation/runAction use pr Read convex_check_type_safety Check Convex code for type safety issues: Read convex_check_validator_coverage Check that all exported Convex functions (query, mutation, action, and their internal variants) have both args Read convex_critter_check The accountability partner that wants to know everything — answer Read convex_discover_tools Search for available convex-mcp-nodebench tools by keyword, category, or task description. Returns matching to Read convex_export_sarif Export stored audit results as SARIF 2.1.0 JSON. SARIF is the industry standard for static analysis results — Read convex_get_methodology Get step-by-step guidance for Convex development workflows. Topics: overview, convex_schema_audit, convex_func Read convex_scan_capabilities Analyze Convex source files for structural patterns. Scans a single file or entire convex/ directory and retur Read convex_search_gotchas Search the persistent Convex gotcha database using full-text search. Returns relevant gotchas, edge cases, and Read convex_snapshot_schema Capture a snapshot of the current schema.ts for later diffing. Stores table names, schema size, and timestamp. Read convex_suggest_indexes Analyze query patterns across all Convex functions and suggest missing indexes. Scans for .query().filter() pa Read convex_verify_concept Check if Convex source files contain all required code signatures for a concept. Provide a concept name (e.g., Read critter_check The accountability partner that wants to know everything — answer Read crypto_price_historical Get historical daily price data for a cryptocurrency. Optionally filter by date range. Read crypto_price_quote Get the latest cryptocurrency price quote. Accepts symbols like BTC, ETH, or BTC-USD. Read csv_aggregate Aggregate values from a local CSV (count/sum/avg/min/max) with optional filters. Deterministic, no network. Read csv_select_rows Select rows from a local CSV using deterministic filters. Returns bounded results (selected columns + matching Read decision_quality_scan Check whether the founder decision has clear criteria, falsifiers, and next actions. Read deeptrace_check_permission Pre-flight permission check — can this agent use this tool? Resolves the agent Read deeptrace_get_evidence_pack Retrieve an evidence pack by ID, including all its constituent chunks with full content and provenance metadat Read deeptrace_get_passport Retrieve an agent Read deeptrace_list_receipts List action receipts with optional filtering by agent, policy action, or tool. Returns receipts sorted by crea Read deeptrace_receipt_stats Get aggregate statistics across all receipts — total count, policy action breakdown, top agents, and top tools Read deeptrace_search_evidence Full-text search across ingested evidence chunks. Returns matching chunks with relevance ranking, source URLs, Read deeptrace_verify_receipt Verify a receipt Read delta_brief Generate a Read delta_compare Side-by-side entity comparison. Produces a delta.diligence comparison packet highlighting differences, strengt Read delta_diligence Deep entity intelligence teardown. Produces a delta.diligence packet with signals, risks, opportunities, and e Read delta_packets List recent delta packets. View your packet history, filter by type, and track lineage. Read delta_review Reconcile a forecast or recommendation against reality. Produces a delta.review packet so the next decision us Read delta_scan Run a self-diligence market coverage scan. Produces a delta.market packet analyzing what NodeBench Delta cover Read delta_watch Add, remove, or list entities on your watchlist. Watched entities are checked for material changes during delt Read design_voice_pipeline Given requirements (latency, privacy, platform, budget), recommend an optimal STT/TTS/LLM stack. Returns top 3 Read detect_company_mode Classify a request as own-company, external-company, or mixed-comparison mode before packet routing. Read detect_contradictions Detect contradictions across multiple sources for an entity. Read detect_repeated_questions Analyze causal_events to find patterns where the user/agent asks the same strategic question repeatedly. Uses Read detect_subvertical Detect the founder subvertical from the query and context. Read detect_temporal_signal Analyze observations in a stream and detect temporal signals: momentum (sustained directional trend), regime_s Read detect_vertical Detect the founder vertical and subvertical from the query and context. Read diff_crawl Before/after site comparison. First call captures baseline. Second call diffs against baseline. Shows added/re Read diff_outputs Compare two text or JSON outputs and produce a structured diff with similarity score. Use for model comparison Read diff_screenshots Compare two images structurally using pixel-level analysis. Computes perceptual similarity, difference regions Read discover_infrastructure Discover existing agent infrastructure in the codebase. Scans for agent loops, telemetry, evaluation, verifica Read discover_tools Multi-modal tool search engine with 14 scoring strategies: keyword, fuzzy (typo-tolerant), n-gram (partial wor Read discover_vision_env Discover available vision-capable AI SDKs and API keys in the current environment. Returns which providers can Read distribution_surface_scan Scan which distribution surfaces are actually ready right now. Read dive_code_locate Find the exact source code location for a bug, component, or design issue. Uses grep/ripgrep to search the pro Read dive_code_review Generate a structured code review report from all dive findings — similar to CodeRabbit or Augment Code Review Read dive_preflight Analyze a project BEFORE starting a UI dive. Scans the project directory to detect: framework (Vite, Next.js, Read dive_snapshot Capture a screenshot or accessibility snapshot of the current page during a dive session. Requires the built-i Read dive_walkthrough Generate a comprehensive page-by-page, component-by-component walkthrough document for a dive session. Include Read economy_gdp Get the latest GDP data for a country from the World Bank. Defaults to the US. Read economy_inflation Get the latest consumer price inflation rate (CPI) for a country from the World Bank. Defaults to the US. Read enrich_entity Enrich an entity (company, person, theme, market) with structured intelligence. Read enrich_recon Retroactively enrich an existing recon session with live web search results. Call this after run_recon when yo Read entity_lookup Quick intelligence on any company, person, or topic. Type a name, get structured facts, signals, and suggested Read equity_fundamental_overview Get a fundamental overview for an equity ticker including price, sector, industry, P/E ratios, market cap, and Read equity_price_historical Get historical daily OHLCV price data for a given equity ticker. Optionally filter by start/end date (YYYY-MM- Read equity_price_quote Get the latest equity/stock price quote for a given ticker symbol. Tries Stooq first, then Yahoo Finance as fa Read evidence_gap_scan List missing evidence classes and materials for diligence readiness. Read export_audit_report Export a session activity log as formatted markdown for pull request reports, Read export_crunchbase_profile Export a Crunchbase-like structured profile from the company packet. Read export_dimension_bundle Export the full DeepTrace dimension bundle for an entity: latest profile, snapshots, evidence, and interaction Read export_pitchbook_profile Export a PitchBook-like structured profile from the company packet. Read export_yc_application_context Export YC-style application context from the company packet. Read exportDocumentToMarkdown Export a document to Markdown format. Returns { title, markdown }. Read extract_figma_frames Extract all frames from a Figma file using depth=3 tree traversal (DOCUMENT -> CANVAS -> SECTION -> FRAME). CR Read extract_fractions_and_simplify_from_image Extract slash-style fractions (e.g. 3/4) from body text in an image and also detect stacked numerator/denomina Read extract_patent_claims Extract likely patent and IP claims from source text. Read extract_publication_metadata Extract publication-oriented metadata from source text. Read extract_regulatory_artifacts Extract regulatory path signals from source text. Read extract_structured_data Extract structured JSON data from unstructured text using an LLM. Provide the text and a description of the fi Read extract_trial_evidence Extract trial, study, or lab evidence snippets from source text. Read extract_variables Identify and weight the key variables driving an entity Read fetch_rss_feeds Fetch and parse all registered RSS/Atom feeds (or specific URLs). New articles are stored in SQLite with is_ne Read fetch_url Fetch a URL and extract its content as markdown, text, or raw HTML. Useful for reading documentation, blog pos Read financial_hygiene_check Return the hidden financial hygiene requirements many founders forget before diligence. Read find_contradictions_for Find all entities that contradict a given entity or concept. Read findTools Search available tools by keyword or capability description. Returns matching tool names and descriptions rank Read forecast_temporal_trend Zero-shot forecasting on numeric time series. Supports naive (last value), linear (regression), and exponentia Read founder_company_naming_pack Generate a founder company naming shortlist and starter profile. Read founder_deep_context_gather MUST be called before generating or updating a Founder Artifact Packet. Read founder_delegation_boundary_scan Separate delegable work from founder-only work for the current direction. Read founder_direction_assessment Pressure-test a founder direction against team shape, AI stance, build speed, Read founder_gaps_detect Detect missing foundations, hidden risks, and weak strategic angles for a founder direction. Read founder_local_gather Gathers all locally-available context for a founder packet: git log, Read founder_materials_check Return the founder materials checklist and missing external-readiness artifacts. Read founder_next_unlocks List the next progression unlocks required to move the founder to the next stage. Read founder_ontrack_scorecard Return explicit 2-week and 3-month on-track or off-track scorecards. Read founder_packet_diff Compares two Founder Artifact Packets and returns a structured diff showing Read founder_packet_history_diff Compares the most recent Founder Artifact Packet for an entity against Read founder_packet_validate Validates a draft Founder Artifact Packet against quality gates before saving. Read founder_readiness_score Return the founder readiness score and a concise interpretation. Read founder_stage_assess Return the founder progression stage, readiness score, and stage ladder for the current direction. Read founder_target_customer_map Map the downstream customer groups the company should target first. Read fusionSearch Multi-source fusion search with reranking. Modes: fast (single source), balanced (2-3 sources), comprehensive Read generate_countermodels For every main thesis or scenario, generate serious alternative explanations with their own evidence and confi Read generate_summary Generate AI-powered summary and insights from completed DCF model. Provides: - Key valuation metrics - Invest Read get_ab_test_report Generate an A/B test comparison report for static vs dynamic toolset loading. Shows session counts, tool count Read get_active_forecasts List active forecasts. Optionally filter by those needing refresh (based on last refresh time and frequency). Read get_agent_role Get the current agent Read get_annual_retrospective Full year view: quarterly summaries, yearly milestones, category distribution, growth metrics. Read get_autopilot_status Check the status of scheduled tasks: when the last run happened, Read get_batch_run_history See the history of past scheduled runs: what was found, Read get_benchmark_history View historical benchmark batch results. Shows RCA and PRR trends over time. Read get_benchmark_oracles Return the oracle definitions for founder autonomy and workflow optimization benchmark lanes. Read get_benchmark_report Get the full detailed report for a specific benchmark batch by batchId. Read get_boilerplate_status Check what NodeBench infrastructure is already set up in a project vs what Read get_causal_chain Trace the causality chain from a given event backwards through causedByEventId links. Returns the chain of eve Read get_compaction_recovery Recovery tool for post-context-compaction state restoration. Call this RIGHT AFTER Claude Code compacts contex Read get_company_truth Get current canonical company truth from subconscious memory blocks. Read get_context_bundle Returns the full NodeBench context bundle: pinned identity (mission, wedge, confidence), Read get_daily_brief_summary Get the latest daily brief summary from local SQLite. Returns dashboard metrics, features, and source summary. Read get_daily_log Get all tracked actions for a specific date, grouped by session with milestone highlights. Read get_design_spec Return the full design governance specification as structured JSON. Includes: approved semantic colors, typogr Read get_design_violations Scan the entire src/ directory for design governance violations. Groups results by file, sorted by severity. S Read get_dimension_profile Fetch the latest persisted DeepTrace dimension profile, including regime label, policy context, confidence, an Read get_distribution_surfaces Inspect NodeBench MCP distribution surfaces: npm/npx, installer, plugin configs, Smithery, and shared web revi Read get_dive_report Generate a comprehensive UI/UX Full Dive report for a session. Includes: executive summary, component tree, al Read get_dive_tree Get the full XML-like component tree for a dive session. Shows all registered components in their hierarchy wi Read get_dogfood_sessions List recent dogfood sessions with their judge scores. Filter by loop type. Read get_dogfood_telemetry Query dogfood telemetry rows with optional filters. Returns matching rows plus computed averages (tool calls, Read get_drift_report Detect configuration and state drift in the NodeBench system. Checks for orphaned verification Read get_engine_context_health Returns accumulated knowledge health: learnings count + freshness, conformance trend direction (improving/stab Read get_entity_graph_summary Get a summary of the knowledge graph: entity counts by type, edge counts by relation, recent additions. Read get_event_ledger Query the causal event ledger with optional filtering by entityId, eventType, entityType, or correlationId. Read get_failure_triage Get open failure cases grouped by system layer with frequency counts. The triage board for fixing system gaps. Read get_figma_design_context Get design context for a specific component or screen from the codebase governance spec. Returns the relevant Read get_flywheel_status Get the current state of both loops (Verification inner loop and Eval outer loop) and how they connect. Shows Read get_forecast_chain Get full audit trail for a forecast: the forecast itself, all evidence entries, update history, and resolution Read get_forecast_evidence Query evidence ledger for a forecast. Optionally filter by signal direction (supporting/disconfirming/neutral) Read get_forecast_track_record Get aggregate Brier scores, calibration summary, and track record statistics across all resolved forecasts. Read get_founder_execution_order Return the canonical founder/company packet execution order so all surfaces follow the same run sequence. Read get_founder_job_topology Return the queue and job topology for founder ingestion, sweeps, deltas, packet refresh, exports, delegation, Read get_founder_packet_resource Fetch the resource URI, pull query, and subscription query for a founder issue or resolution packet. Read get_founder_progression_rubric Return the explicit founder progression rubric, including mandatory and optional signals for each stage. Read get_gate_history Get the history of quality gate runs for a given gate name. Shows pass/fail trend over time. Read get_gate_preset Get the rules for a built-in quality gate preset. Returns rule names, descriptions, and evaluation instruction Read get_important_changes Query flagged important changes with optional status filtering. Returns changes ordered by timestamp descendin Read get_improvement_recommendations Analyze all persisted data and return actionable improvement recommendations. Detects: unused tools, missing q Read get_ingest_status Report which data sources are available for ingestion. Checks if Claude Code transcripts exist, what project d Read get_judge_history Get the history of LLM judge runs, optionally filtered by scenario. Read get_latest_signals Get the most recent signal sweep results without running a new sweep. Returns cached signals from the last swe Read get_messaging_health Check which messaging channels are working. Read get_monthly_report Monthly rollup of actions: weekly breakdown, trending categories, velocity (actions/day), milestone timeline. Read get_narrative_status Get narrative thread status from local SQLite. Returns threads grouped by phase (emerging, escalating, climax, Read get_observability_summary Unified observability summary combining MCP system pulse, sentinel probes, watchdog status, Read get_openclaw_audit View the activity log for OpenClaw sessions. Shows every skill call Read get_openclaw_delivery_status Check if your messages were delivered. Read get_openclaw_results Get results and safety summary for an OpenClaw session. Read get_ops_dashboard Get operational dashboard status from local SQLite. Returns last sync info, tool call frequency, active verifi Read get_packet_lineage Trace the full derivation chain for a packet or entity. Read get_parallel_status Get a comprehensive overview of all parallel agent activity: active task claims, role assignments, context bud Read get_policy_detail View full details of a sandbox policy including effective allowlist/blocklist after merging defaults. Read get_proactive_alerts Scan causal memory for watchlist-worthy alerts: new events on tracked entities, unresolved important changes, Read get_project_context Retrieve the stored project context (tech stack, architecture, conventions, etc.) and knowledge base stats. Ca Read get_quarterly_review Quarterly strategic view: monthly trends, category shifts, biggest milestones, velocity curve. Read get_recon_summary Get aggregated summary of all findings from a recon session. Groups findings by category (breaking changes, ne Read get_regression_gate Check if the 3 canonical loops pass. Returns per-loop scores, overall pass/fail, and regression detection. Read get_repeat_cognition_metrics The key compound metric. Measures repeat question rate, manual reconstruction count, packet abandonment rate, Read get_role_packet_defaults Return the default packet, artifact, monitor, and delegation policy for a specific public role lens. Read get_self_directed_delivery_run Load a previously recorded autonomous delivery run with all stage receipts, summaries, and final recommendatio Read get_self_eval_report Generate a comprehensive self-evaluation report by cross-referencing all persisted data: verification cycles, Read get_sentinel_report Get the latest sentinel self-test report with all 9 probe results (build, e2e, design, dogfood, Read get_session_compliance Score an OpenClaw session against its security policy. Read get_session_journal Get all tracked actions from the current or a specified session, in chronological order. Read get_session_profile Get the efficiency profile for the current session. Shows total calls, cost, latency, redundancy, and optimiza Read get_shared_context_packet Fetch one shared-context packet directly and include the suggested resource URI plus pull/subscription filters Read get_shared_context_peer Fetch one peer and its current summary, capabilities, and scopes. Read get_shared_context_snapshot Inspect the current shared-context protocol state: peers, packets, handoffs, messages, and aggregate counts. Read get_signal_recommendations Get founder-specific actionable recommendations from the latest signals. Each recommendation includes: what to Read get_skill_risk_profile Get the safety assessment for one or more OpenClaw skills: Read get_source_trust_policy Return the source-level permission and trust policy for storage, summarization, and export across private and Read get_state_diff_history Get the change history for a specific entity, showing all recorded state diffs in reverse chronological order. Read get_subconscious_hint Get the subconscious Read get_sync_bridge_status Inspect whether the device is local-only, connected but idle, or actively ready to sync to a paired web accoun Read get_system_pulse Get a real-time health snapshot of the NodeBench MCP system. Returns database status, Read get_tool_graph Returns a JSON graph of tool relationships that grows with usage. Read get_tool_quick_ref Get the quick reference for a specific tool: what to do next, related tools, methodology, and tips. Call this Read get_trajectory_analysis Analyze tool usage trajectories across sessions. Returns tool frequency, error rates, average duration, phase Read get_trajectory_summary Compute a trajectory summary for a date range: event counts by type, diff counts by change type, path step cou Read get_traversal_plan Generate a traversal plan for accomplishing a goal across multiple views. Read get_uptime_stats Get session uptime, tool call rates, error trends, and top tools over multiple time windows Read get_usage_insights Get aggregate usage insights across all sessions. Shows top tools, repeated queries, cost breakdown, and optim Read get_verification_status Get the current status of a verification cycle including all phases, gaps, and test results. Use this to under Read get_view_capabilities Get full capabilities for a specific view — actions, data endpoints, Read get_view_state Get the current agent traversal session state — which view you Read get_watchdog_log Get recent watchdog check results. Shows health score trend, detected issues, and auto-healed Read get_weekly_summary Summarize a week Read get_workflow_chain Get a recommended tool sequence for a common workflow. Returns step-by-step tool chain with actions for each s Read get_workflow_history Returns past runs for a specific workflow with scores, grades, step counts, and durations. Use to track confor Read getArtifactsWithHealth Get source artifacts annotated with verification health status. Shows which sources are verified, stale, or co Read getCalibrationStats Get calibration statistics for the verification pipeline, including accuracy, F1 scores, and inter-annotator a Read getCanaryRegressions Get recent canary runs that detected regressions. Read getCanaryTrend Get canary benchmark trend: throughput score, quality score, verdict, and delta over time. Read getComparisonTrend Get baseline comparison trend for a benchmark family (investigation, canary, etc.). Read getContradictoryPosts Get posts that contradict each other across threads, useful for fact-checking and dispute resolution. Read getDealFlow Get the current deal flow pipeline with funding events, company data, and investment signals. Read getDocument Get a document by ID, including title, content, metadata, and file info. Read getEntity Get a canonical entity by its entityKey. Read getEntityContext Get full entity context by ID, including all tracked metadata, relationships, funding data, and activity histo Read getEntityContextByName Look up an entity context by its canonical name (exact match). Read getEntityContextStats Get aggregate statistics about the knowledge base: total entities, entities by type, coverage metrics. Read getEntityInsights Get deep entity insights for a company or person, including funding, key people, product pipeline, and persona Read getEntityNetwork Get the full network graph for an entity: aliases, holdings, people, investors, and related entities. Read getEvolutionDashboard Get the evolution dashboard: latest canary, pending routing recommendations, recent comparisons, and error sum Read getFactById Get a single fact by ID with its full context: source, extracted claim, linked artifacts, and verification sta Read getFactsByRun Get all extracted facts from a specific verification run (e.g., from a daily brief or narrative pipeline run). Read getFastestGrowingRepos Get fastest-growing GitHub repositories by star velocity over the past week. Read getFolderWithDocuments Get a folder and all documents inside it. Read getForYouFeed Get the personalized For You feed with ranked research items, funding events, and industry signals. Returns ve Read getHoldings Get all fund holdings for a given investor entity. Read getHygieneReport Get post-execution hygiene report: stale missions, pending sniff checks, error rates. Read getInvestorProfile Get investor profile with type, AUM, focus sectors, portfolio size. Read getKnowledgeGraph Get a knowledge graph by source document, including nodes, edges, and claims. Read getKnowledgeGraphClaims Get all claims extracted from a knowledge graph, useful for fact-checking and verification. Read getLatestDashboard Get the latest research dashboard snapshot with metrics on deal flow, entity coverage, verification health, an Read getLatestPublicDossier Get the latest public company/industry dossier with competitive analysis and market positioning. Read getMethodology Get step-by-step guidance for a development methodology. Topics: verification, eval, flywheel, mandatory_flywh Read getMigrationStats Get model migration statistics for chat threads and messages. Read getMission Get a specific mission by its Convex ID. Read getMissionByKey Look up a mission by its unique missionKey string. Read getMissionDashboard Get the mission control dashboard: active missions, judging queue, sniff checks, and recent completions. Read getModelCostSummary Get cost, token, latency, and error summary for a specific model over N days. Read getOpenDisputes Get all open disputes (contradictions, contested facts) that need resolution across the narrative engine. Read getPendingRoutingRecommendations Get pending model routing recommendations from the eval engine. Read getPendingSniffChecks Get all pending human-in-the-loop sniff checks awaiting review. Read getPersonProfile Get executive/person profile with employment history, board roles, and credibility score. Read getPlan Retrieve a task plan by ID Read getPublicThreads List all public narrative threads. Each thread tracks an evolving story across entities over time (e.g., Read getSignalTimeseries Get time-series signal data for research metrics like funding volume, entity mentions, or verification health Read getSloMetricsSummary Get SLO metrics for verification pipeline performance: latency percentiles, throughput, error rates. Read getSourceRegistry Get the source registry for a domain, showing all tracked data sources with freshness and reliability scores. Read getSpreadsheetRange Get cells within a range from a spreadsheet (row/col coordinates). Read getStakeholderGraph Get the stakeholder analysis graph: actors, goals, fears, incentives, and incentive interpretations. Read getTasksForMission Get all task plans for a mission, showing execution status, dependencies, and assigned agents. Read getThread Get a single narrative thread by ID, including metadata, entity associations, and event count. Read getThreadPosts Get posts (analyst notes, thesis updates, evidence) within a narrative thread. Read getThreadsByEntity Get all narrative threads associated with a specific entity (company or person name). Read getThreadStats Get aggregate statistics about the narrative engine: total threads, events, posts, disputes, and activity over Read getThreadsWithEvents Get narrative threads with their recent events inlined, useful for getting a quick overview of active stories. Read getTrendingRepos Get trending GitHub repositories tracked by NodeBench AI, sorted by growth velocity. Read getVerificationsForFact Get all verification attempts for a specific fact, including source citations, verdicts, and confidence scores Read getVerificationSummary Get a summary of verification health across all tracked facts. Returns counts by verdict (VERIFIED, CORROBORAT Read grade_agent_run Grade a single agent run on both outcome quality (task success, regressions, time) and process quality (recon, Read grade_fraction_quiz_from_image Grade a fraction quiz shown in an image by OCRing the problems + student answers, computing correct answers, a Read graphify_report Get the GRAPH_REPORT.md analysis from a graphify run. Contains god nodes (most connected), Read graphify_status Check if graphify is installed and ready. Returns version, installation instructions if missing. Read heartbeat_shared_context_peer Refresh peer liveness and optionally publish a compact machine-readable summary of its current work. Read hiring_gap_scan Identify the most obvious missing hiring lane for the current founder direction. Read ingest_claude_code_sessions Scan Claude Code JSONL transcripts from ~/.claude/projects/. Returns session summaries with turn counts, token Read ingest_codebase_changes Fingerprint key files in a directory (package.json, README, schema, CLAUDE.md, etc.) and detect what changed s Read inject_context_into_prompt Wraps a user prompt with NodeBench Read inspect_topology_shape Inspect the persisted Convex topology shape for the Reports graph. Use this after search_memory/search_report_ Read invalidate_shared_context Invalidate a packet when it becomes stale, contradicted, or superseded. Read investigate Investigate a company, person, or topic and return a concise sourced artifact. This is the default v3 entry po Read json_select Select a sub-value from a local JSON file using a JSON Pointer (RFC 6901) and return a bounded preview. Determ Read judge_verify_subtask Judge verifies a subtask Read list_agent_tasks List all current task claims across parallel agents. Shows who is working on what, blocked tasks, and recently Read list_available_toolsets List all available toolsets showing which are currently loaded and which can be dynamically added. Includes to Read list_available_views List all available views in the NodeBench AI frontend with titles, Read list_contradictions List all open contradictions between memory blocks and graph entities. Read list_dimension_evidence List the durable evidence rows behind a DeepTrace dimension profile. Useful for auditing why a score or availa Read list_dimension_interactions List stored interaction effects for an entity, such as capital plus investor quality reducing execution fragil Read list_dimension_snapshots List historical DeepTrace dimension snapshots for an entity to inspect regime transitions over time. Read list_driver_tools List all tools available from connected MCP drivers. Shows tool names, descriptions, and input schemas. Use th Read list_eval_runs List recent eval runs with their aggregate scores. Use this to track quality over time and detect drift. Read list_extracted_skill_templates List reusable skill templates automatically extracted from successful harness runs. Read list_founder_issue_packets List founder issue packets from shared context by workspace, producer, status, or strategic angle. Read list_learnings [DEPRECATED: Use search_all_knowledge instead] List stored learnings. PREFER search_all_knowledge for unified Read list_openclaw_channels List messaging channels you can send through. Read list_pending_sync_operations List queued outbound sync operations that still need to be pushed to the web account. Read list_sandbox_policies List all configured sandbox policies with session stats (active sessions, total violations triggered). Read list_self_directed_delivery_runs List recent autonomous delivery runs so operators can reopen or compare them. Read list_shared_context_peers List peers by product, workspace, role, surface, capability, or scope. Read list_skills List all registered skills with their freshness status, source files, Read list_stale_packets List memory blocks and packets that need refresh. Read list_verification_cycles List all verification cycles, optionally filtered by status. Use this to find a cycle ID or review past verifi Read list_webmcp_tools List all tools discovered from connected WebMCP origins. Shows tool names, descriptions, and input schemas. Op Read list_workspace List files in the agent workspace. Shows folder tree with file sizes and dates. Read listAgentMemory List all stored memory keys with their metadata Read listDocuments List all user documents sorted by last modified (most recent first, up to 200). Read listEntityContexts List all tracked entity contexts with pagination. Useful for browsing the knowledge base. Read listFiles List user files, optionally filtered by type (image, video, document, csv, pdf). Read listFolders List all user folders. Read listSpreadsheets List user spreadsheets sorted by last updated. Read load_diligence_pack Load the vertical diligence pack for the current direction. Read load_session_notes Load session notes from the filesystem. Use after context compaction, /clear, or session resume to recover sta Read log_context_budget Track context window usage to prevent pollution. LLM agents have finite context and, as Anthropic Read log_tool_call Log an MCP tool call for profiling. Called automatically by the NodeBench gateway to track tool usage patterns Read lookup_entity Look up a single entity by URI (entity://<slug>) or by canonical name. Backed by the federated search action f Read meeting_notes_extract_decisions Extract decisions, owners, and follow-ups from raw meeting notes. Read mine_session_patterns Analyze tool_call_log and learnings tables to extract recurring success/failure sequences across sessions. Fin Read monitor_repo Track a GitHub repository Read multi_criteria_score Deterministic weighted multi-criteria decision analysis (MCDM). Takes options with numeric values per criterio Read nb_filter_receipts Filter action receipts by channel, policy, approval state, or direction. Read nb_filter_tool_activity Filter the MCP ledger by tool name, date, or status. Read nb_get_feed_items Get public research feed items surfaced on Home. Read nb_get_funding_brief Get funding intelligence through the active research surface. Read nb_get_leaderboard Get model evaluation leaderboard data for the internal benchmark surface. Read nb_get_qa_results Get the latest QA pipeline results, scores, issues, and governance violations. Read nb_get_signals Get the latest research signals with trend data and source attribution. Read nb_list_agents List available AI agent templates and active conversations. Read nb_list_escalated_actions List escalated actions that triggered policy review or human approval. Read nb_list_scenarios List available eval scenarios with descriptions and difficulty. Read nb_open_delegation Open the delegation surface to review scoped authority and approval gates. Read nb_open_receipts Jump into the live receipts feed from the home surface. Read nb_review_pending_approvals Review approval-gated receipts waiting for a human decision. Read nb_review_shared_context Inspect shared-context peers, packets, and handoffs. Read nb_search_from_chat Run a grounded search from the chat surface. Read nb_search_reports Search saved reports and public packets. Read nb_search_research Search across research signals, briefings, and intelligence reports. Read nb_view_screenshots Get captured route screenshots for visual QA. Read news_company Get recent news articles for a specific company/ticker symbol from Yahoo Finance. Read news_world Get world/market news headlines. Currently returns a placeholder — requires an external news API for full func Read nodebench_get_app_state Get the current application state — which view is active, its capabilities (data endpoints, actions, tools), a Read nodebench_get_capabilities Meta-tool: lists all available WebMCP tools with their descriptions. Call this first to understand what you ca Read nodebench_get_digest Retrieve the latest morning digest — a curated summary of research signals, trending topics, and actionable in Read nodebench_get_screen_context Get the current screen Read nodebench_list_views List all available views with their capabilities — titles, descriptions, actions, data endpoints, and tags. Us Read nodebench_query_elements Query all agent-annotated interactive elements on the current page. Returns elements with their data-agent-id, Read nodebench_search Search NodeBench documents, research dossiers, and knowledge base. Returns relevant results with titles, snipp Read nodebench_search_views Search for views by keyword — matches against titles, descriptions, tags, and tool categories. Use this to fin Read nodebench_traverse_feed Traverse content feeds with Moltbook-style sorting (hot, new, top, rising). Supports pagination via cursor. Av Read nodebench.activity_timeline Read the canonical activity ledger for a NodeBench report, including captures, notebook patches, graph clicks, Read nodebench.compile_interview_packet Compile a public interview-prep context pack for a company or person. Resume-specific fit scoring must happen Read nodebench.context.pack Return a compact context pack for a use case like job_match, interview_prep, sales_research, or product_intel. Read nodebench.dossiers.get Get a sourced public dossier for an entity by entity ID, key, name, or domain. Read nodebench.entities.resolve Resolve a public company, person, role, product, investor, school, or source into a stable NodeBench entity ID Read nodebench.expand_resource Expand a Nodebench resource URI (nodebench://org/{key}) by one ring using the requested lens + depth. Returns Read nodebench.expand_subject Expand a single person, company, event, topic, repo, or document into canonical entity resolution, related ent Read nodebench.get_entity_dossier Get a sourced public entity dossier from NodeBench public research memory by entity_key or name/type. Read nodebench.get_matching_context Alias for nodebench.context.pack optimized for app-local matching and scoring flows. Read nodebench.refresh_resource Refresh a stale Nodebench resource URI and return updated metadata, freshness information, and emitted resourc Read nodebench.render_output Render an existing brief or artifact into a target format such as compact alert, ntfy brief, Notion markdown, Read nodebench.report_export_preview Prepare a reviewable NodeBench report export. Returns mapped contacts, companies, interactions, follow-ups, cl Read nodebench.research_company Run or reuse public-source company research, store verified public claims, and return a compact company dossie Read nodebench.research_person Run or reuse public-source person research, store verified public claims, and return a compact person dossier. Read nodebench.search_public_sources Search NodeBench public research memory and source cache for public facts about an entity. Private email, resu Read open_operating_dashboard Start the Operating Dashboard — shows trajectory scores, event ledger, important changes, path replay, time ro Read overstory_fleet_status Read the Overstory multi-agent fleet status — active agents, capabilities, health, worktree state. Read overstory_mail_log Read recent Overstory QA mail messages — capture-complete, stability results, Read overstory_merge_queue Read the Overstory merge queue — pending merges, completed merges, Read overstory_qa_summary Aggregate QA gate results across all routes — per-route stability grades, Read partnership_target_map Map likely partnership targets and why they fit the current wedge. Read pdf_search_text Search text inside a local PDF over selected pages. Returns page numbers and bounded snippets around matches. Read populate_historicals Populate historical financial data from discovered sources. This tool: - Checks NodeBench cache first (instan Read predict_risks_from_patterns Given a task description, predict likely failure modes based on historically similar sessions. Searches the le Read pull_profile Pull the current company profile and saved context from NodeBench AI into Claude Code. Returns the latest shar Read pull_report Pull a saved report from NodeBench AI into Claude Code. Returns the full report with sections, sources, and ne Read pull_shared_context Pull shared-context packets by type, producer, scope, workspace, status, or subject substring. Read query_daily_brief Get today Read query_funding_entities Search funding intelligence from the Convex platform. Filter by company name, round type, or get recent events Read query_graphify Query an existing graphify knowledge graph. Find nodes by label, explore connections, Read query_research_queue View the research task queue from the Convex platform. Shows active and pending research topics with prioritie Read query_temporal_signals Search and retrieve temporal signals with filtering by entity, signal type, status, and date range. Returns fo Read query_view_data Query data from a view Read quickSearch Execute a quick multi-source search query. Returns ranked results from web, news, and academic sources. Read rank_interventions Rank potential interventions by expected trajectory delta. Each intervention includes expected impact, confide Read read_csv_file Read a local CSV file and return a bounded table preview (headers + rows). Deterministic, no network. Read read_docx_text Extract text from a local DOCX (Office OpenXML) file. Deterministic, no network. Read read_emails Read emails from an IMAP mailbox over TLS. Requires EMAIL_USER and EMAIL_PASS env vars. Defaults to Gmail IMAP Read read_image_ocr_text Extract text from a local image (PNG/JPG/etc) using OCR (tesseract.js). Deterministic, no network. Read read_json_file Read a local JSON file and return a bounded JSON preview (depth/item/string truncation). Deterministic, no net Read read_jsonl_file Read a local JSONL file and return bounded parsed rows. Deterministic, no network. Read read_pdf_text Extract text from a local PDF file for selected pages. Returns bounded text with page markers. Deterministic, Read read_pptx_text Extract text from a local PPTX (Office OpenXML) file. Deterministic, no network. Read read_text_file Read a local text file (txt/md/xml/json/etc) and return a bounded text slice. Deterministic, no network. Read read_workspace_file Read a file from the agent workspace. Returns content for text files, metadata for media files. Read read_xlsx_file Read a local XLSX workbook and return a bounded sheet preview (headers + rows). Deterministic, no network. Read readAgentMemory Retrieve previously stored data by key Read refresh_subconscious Get a summary of all subconscious memory blocks and their status. Read refresh_task_context Re-inject the current task context to combat attention drift. After 30+ tool calls, models lose sight of origi Read render_decision_memo Render a 1-page executive decision memo from a completed Deep Sim analysis. Combines claim graph, variables, s Read research_job_market Research job market requirements for a given role or skill set. Provides guidance on in-demand skills, common Read resolveAlias Resolve an alias (name, ticker, CUSIP, ISIN) to its canonical entity. Read retention_get_status Return the latest retention.sh connection and recent event history from local MCP state. Read retention_status Check retention.sh connection status and QA metrics. Shows team code, QA score, member count, and last sync ti Read review_paper_as_reviewer Simulate a peer reviewer evaluating a paper for a top venue. Default: harsh mode with rejection mindset — only Read review_pr_checklist Structured PR review checklist with verification/eval cross-reference. Validates PR title format, description Read route_founder_packet Route a founder/company request into the canonical company mode, packet type, artifact type, and next action p Read runway_check Basic runway check that translates cash and burn into months remaining and flags risk. Read sandbox_search BM25-ranked full-text search across all sandboxed content. Pass multiple queries as an array to batch all ques Read sandbox_stats Show context savings for the current session — per-tool breakdown, total bytes indexed vs returned, savings ra Read scan_capabilities Analyze a source file for structural code patterns. Returns a capability report showing what the code can stru Read scan_dependencies Scan a project Read scan_terminal_security Scan project files and dev environment for terminal security threats: Unicode homograph attacks, ANSI escape i Read scan_webmcp_origin One-shot scan: connect to a WebMCP-enabled site, discover tools, cache the manifest, and disconnect. Useful fo Read scan_wordpress_updates Scan a WordPress site for plugin and theme versions, and optionally check for known vulnerabilities via the WP Read score_compounding Compute the full 8-dimension trajectory score for an entity. Returns trust-adjusted compounding, drift, adapta Read score_scenario_branch Score a specific scenario branch against evidence and constraints. Read scrapling_batch_fetch Fetch multiple URLs in parallel with configurable concurrency. Use for competitive analysis, multi-source rese Read scrapling_crawl_status Check crawl progress and get collected items. Pass the session_id from scrapling_crawl. Returns status (runnin Read scrapling_extract Extract structured data from a URL using CSS or XPath selectors. Zero LLM tokens — deterministic extraction. U Read scrapling_fetch Fetch a URL with adaptive scraping. Auto-selects fetcher tier: Read scrapling_track_element Track an element across page versions using Scrapling Read search Search across live web results and stored NodeBench knowledge in one call. Use this instead of deciding betwee Read search_all_knowledge Search ALL accumulated knowledge in one call: learnings (edge cases, gotchas, patterns), recon findings (acros Read search_content_archive Search past content by theme, title, or keywords using FTS5 full-text search. Use before generating new conten Read search_github Search GitHub repositories by query, topic, language, and star count. Useful for discovering libraries, framew Read search_learnings [DEPRECATED: Use search_all_knowledge instead] Search past learnings. PREFER search_all_knowledge which search Read search_memory Federated keyword search across NodeBench Read search_openclaw_gotchas Search stored OpenClaw pitfalls, security findings, and best practices. Read search_report_context Search within a single report Read searchDocuments Search documents by title (full-text search). Read searchEntities Search entities by type, sector, or country. Returns up to 50 results. Read searchEntityContexts Search entity contexts by keyword. Returns entity profiles with summaries, tags, sectors, and relationship dat Read searchResearch Search across research content Read searchThreads Full-text search across narrative threads by keyword. Read seo_audit_url Fetch a URL and analyze its SEO elements: title tag, meta description, Open Graph tags (og:title, og:descripti Read service_to_dashboard_path Map a concept from bespoke service work to a possible dashboard subscription path without losing the local-fir Read setup_local_env Discover and diagnose the local development environment. Checks for available API keys, installed SDKs, Node.j Read share_get_packet_link Retrieve a local share link record by share ID. Read shortest_valid_path Return the shortest valid workflow path for the stated objective. Read solve_bass_clef_age_from_image Extract bass-clef note letters from a simple staff image and compute the derived Read solve_green_polygon_area_from_image Compute the area of a green filled polygon in an image by pixel segmentation, calibrating pixel-to-unit scale Read solve_red_green_deviation_average_from_image Extract red and green numbers from an image, compute population stdev(red) and sample stdev(green), then retur Read solve_storage_upgrade_cost_per_file_from_image OCR plan tiers from an image, compute required storage from equally-sized file counts, and return average incr Read submission_readiness_score Score whether the company packet is ready for downstream submission or profile export. Read suggest_optimizations Analyze the current session and suggest specific optimizations: cheaper models, cached results, workflow short Read suggest_related Suggest related items given a root URI (entity://, report://, etc.). PR1 implementation uses the root Read suggest_tests Generate scenario-based test suggestions from a site_map or diff_crawl session. Analyzes crawl findings and ge Read summarize_session Summarize the current or specified session Read sync_figma_tokens Pull Figma design token variables from a Figma file and compare against the codebase Read team_alignment_check Check whether the team is aligned on the wedge, next move, and moat story. Read thompson_anti_elitism_lint Scan content for elitism, gatekeeping language, and intellectual intimidation. Deterministic banned-phrase det Read thompson_feynman_edit Skeptical Beginner editor — reviews Thompson-written content against 8 rejection criteria. Returns PASS/REWRIT Read thompson_quality_gate Deterministic 10-point quality gate for Thompson Protocol content. Produces a boolean checklist and overall gr Read thompson_visual_map Generate precise visual prompts that map 1:1 with content analogies. No generic b-roll — every visual reinforc Read toon_decode Convert TOON (Token-Oriented Object Notation) string back to JSON. Use this to parse TOON-encoded data from ot Read toon_encode Convert JSON data to TOON (Token-Oriented Object Notation) format. TOON uses ~40% fewer tokens than JSON while Read track_entity_changes Detect what changed for an entity since a given date. Read translate_academic Translate academic text between Chinese and English, preserving LaTeX commands, citations, equations, and tech Read traverse_entity_graph Find all entities and relationships connected to a starting entity within N hops. Read traverse_feed Traverse content feeds with Moltbook-style sorting. Read validate_shortcut Validate that a proposed shortcut preserves output quality and visibility. Read verify_concept_support Check if a source file contains all required code signatures for a concept. Provide a concept name and a list Read watchlist_get_alerts Return watchlist entries with attached change summaries or non-zero alert counts. Read watchlist_list_entities List watched entities from the local founder watchlist. Read web_search Search the web through NodeBench Read workflow_adoption_scan Evaluate how naturally a direction fits current high-frequency user workflows, install surfaces, and maintenan Read xlsx_aggregate Aggregate values from a local XLSX (count/sum/avg/min/max) with optional filters. Deterministic, no network. Read xlsx_select_rows Select rows from a local XLSX using deterministic filters. Returns bounded results (selected columns + matchin Read zip_list_files List entries in a local ZIP file. Deterministic, no network. Read zip_read_text_file Read a text file inside a local ZIP archive and return bounded text. Deterministic, no network.
EXPLORE

Related servers

Other MCP servers with similar tools — same risk classification, starter policies for each.

UnClick 1658 tools
adminautomation
Yaver 811 tools
adminautomation
Binance MCP Server 734 tools
adminautomation
Crow 587 tools
adminautomation
Fortimanager 584 tools
adminautomation
TinyFn 572 tools
adminautomation
Mcp 571 tools
adminautomation
GoHighLevel MCP Server 566 tools
adminautomation
FAQ

Questions about Nodebench

Can an AI agent delete data through the Nodebench MCP server? +

Yes. The Nodebench server exposes 10 destructive tools including abandon_cycle, archiveDocument, cleanup_stale_runs. These permanently remove resources with no undo. PolicyLayer blocks destructive tools by default so they never reach the upstream server.

How do I prevent bulk modifications through Nodebench? +

The Nodebench server has 192 write tools including accept_shared_task, ack_shared_context, add_forecast_evidence. Set a rate limit in your policy -- for example, 10 calls per hour prevents an agent from making more than 10 modifications per hour. PolicyLayer enforces this at the gateway, before calls reach Nodebench.

How many tools does the Nodebench MCP server expose? +

824 tools across 4 categories: Destructive, Execute, Read, Write. 512 are read-only. 312 can modify, create, or delete data.

How do I enforce a policy on Nodebench? +

Register the Nodebench MCP server in PolicyLayer, apply the suggested rules above (adjust the limits to your use case), and point your AI client at the PolicyLayer proxy URL instead of the server directly. Your agents keep the same tools; PolicyLayer evaluates every call against policy before it executes. Nothing to install, live in minutes.

Enforce policy on every Nodebench tool call.

Deterministic rules across all 824 Nodebench tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON NODEBENCH →

Instant setup, no code required.

824 Nodebench tools catalogued and risk-classified — across an index of 46,500+ MCP servers.

// WHERE THIS COMES FROM

These policies come from Nodebench's registry record.

The record behind this page: verified identity, auth posture, risk grade, every tool classified, recommended policy — re-checked continuously.

Teams ship this data inside their own products. See what a licence covers →

// GET IN TOUCH

Have a question or want to learn more? Send us a message.

Message sent.

We'll get back to you soon.