High-risk tools in Ruflo

Manage a fixed-size warm pool of pre-spawned agents to skip cold-start cost on bursty workloads. Use when native Task is wrong because (a) you have a queue of similar tasks and ...

Spawn a subagent

Run memory consolidation to promote entries across tiers and compress old data Use when generic memory_* tools are wrong because you need AgentDB-specific controllers (HNSW vect...

Route a task via AgentDB SemanticRouter or LearningSystem recommendAlgorithm Use when generic memory_* tools are wrong because you need AgentDB-specific controllers (HNSW vector...

Start a session with ReflexionMemory episodic replay Use when generic memory_* tools are wrong because you need AgentDB-specific controllers (HNSW vector search, hierarchical ti...

Disable autopilot. Agents will be allowed to stop even if tasks remain. Use when running long-horizon goals that should resume automatically across sessions — Claude Code has no...

Enable autopilot persistent completion. Agents will be re-engaged when tasks remain incomplete. Use when running long-horizon goals that should resume automatically across sessi...

Discover success patterns from past task completions. Requires AgentDB for full functionality. Use when running long-horizon goals that should resume automatically across sessio...

Predict the optimal next action based on current state and learned patterns. Use when running long-horizon goals that should resume automatically across sessions — Claude Code h...

Execute performance benchmarks (V2 compatible). Deprecated: Use system/metrics instead.

Navigate back in browser history Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay agains...

Check a checkbox Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-drifted ve...

Click an element using ref (@e1) or CSS selector Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reus...

Execute JavaScript in page context Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay agai...

Clear and fill an input element Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against...

Navigate forward in browser history Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay aga...

Hover over an element using ref (@e1) or CSS selector Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie...

Navigate browser to a URL Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-d...

Press a keyboard key Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-drifte...

Reload the current page Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-dri...

Scroll the page Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-drifted ver...

Select an option from a dropdown Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay agains...

End a recorded browser session: trajectory-end with verdict, rvf compact, AIDefence pre-store gate (best-effort), and AgentDB index in the browser-sessions namespace. Use when n...

Open a named, traced browser session: allocate an RVF cognitive container, begin a ruvector trajectory, then open the URL via agent-browser. Returns the session id and rvf path....

Fetch a recipe from the browser-templates AgentDB namespace and return it for caller-level execution. Use when native WebFetch is wrong because you need real browser automation ...

Type text with key events (for autocomplete, etc.) Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie re...

Uncheck a checkbox Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-drifted ...

Wait for a condition Use when native WebFetch is wrong because you need real browser automation — JS-heavy SPA scraping, login flows with cookie reuse, replay against DOM-drifte...

ADR-164 Phase 3 — Compute the domain-affinity routing decision for a business pod per ADR-164 §3.4 and return {backend, reason}. The three backends are local-stdio (preferLocalE...

Accept a pending handoff Use when nothing native covers per-agent capability gating — Claude Code agents have file-system access by default. Pair claims_grant + claims_check bef...

Request handoff of an issue to another claimant Use when nothing native covers per-agent capability gating — Claude Code agents have file-system access by default. Pair claims_g...

Suggest or apply load rebalancing across agents Use when nothing native covers per-agent capability gating — Claude Code agents have file-system access by default. Pair claims_g...

Steal a stealable issue Use when nothing native covers per-agent capability gating — Claude Code agents have file-system access by default. Pair claims_grant + claims_check befo...

Request collective decision from agents

Manage consensus protocol with BFT, Raft, or Quorum strategies Use when native Task is wrong because the work crosses multiple agents that need to vote/sync/load-balance — TodoW...

Configure load balancing Use when native Task is wrong because the work crosses multiple agents that need to vote/sync/load-balance — TodoWrite + a single Task cannot orchestrat...

Manage coordination nodes Use when native Task is wrong because the work crosses multiple agents that need to vote/sync/load-balance — TodoWrite + a single Task cannot orchestra...

Orchestrate multi-agent coordination Use when native Task is wrong because the work crosses multiple agents that need to vote/sync/load-balance — TodoWrite + a single Task canno...

Synchronize state across nodes Use when native Task is wrong because the work crosses multiple agents that need to vote/sync/load-balance — TodoWrite + a single Task cannot orch...

Trigger agent adaptation based on feedback Use when native Task is wrong because you need agents that adapt their cognitive pattern (convergent / divergent / lateral / systems /...

Create a decentralized autonomous agent Use when native Task is wrong because you need agents that adapt their cognitive pattern (convergent / divergent / lateral / systems / cr...

Execute a DAA workflow Use when native Task is wrong because you need agents that adapt their cognitive pattern (convergent / divergent / lateral / systems / critical) per-task ...

Initialize the ONNX embedding subsystem with hyperbolic support Use when text similarity matters beyond keyword match — native Grep finds exact strings, embeddings find meaning....

Build RaBitQ 1-bit quantized index from stored embeddings (32× compression). Pre-filters candidates via Hamming scan before exact rerank. Use when text similarity matters beyond...

Emit an event to subscribers

Execute a workflow

Propose a federated consensus operation across all active peers

Initialize federation on this node with a manifest and begin discovery

Send a message to a federated peer through the PII pipeline and security gates. Optional budget controls (ADR-097): maxHops defaults to 8 to prevent recursive delegation; maxTok...

Manage GitHub Actions workflows Use when native Bash / file tools are wrong because this MCP tool exposes Ruflo-specific state or controllers that have no shell equivalent. For ...

Cook a formula into a protomolecule with variable substitution (WASM accelerated)

Execute a formula to create beads/molecules in Gas Town

Batch cook multiple formulas using WASM for maximum performance

Optimize convoy execution order using WASM graph algorithms

Parse TOML formula content to AST using WASM (352x faster than JavaScript)

Resolve bead dependencies using WASM (topological sort, cycle detection, critical path)

Broadcast message to all workers Use when native Task is wrong because you need queen-led collective intelligence — Byzantine-FT consensus, broadcast across many worker agents, ...

Propose or vote on consensus with BFT, Raft, or Quorum strategies Use when native Task is wrong because you need queen-led collective intelligence — Byzantine-FT consensus, broa...

Initialize the hive-mind collective Use when native Task is wrong because you need queen-led collective intelligence — Byzantine-FT consensus, broadcast across many worker agent...

Join an agent to the hive-mind Use when native Task is wrong because you need queen-led collective intelligence — Byzantine-FT consensus, broadcast across many worker agents, sh...

Spawn workers and automatically join them to the hive-mind (combines agent/spawn + hive-mind/join) Use when native Task is wrong because you need queen-led collective intelligen...

Generate optimized agent configurations from pretrain data Use when native Bash hooks (via Claude Code\

Apply a deterministic, $0 (no-LLM) code transform — the real Tier-1 execution path (ADR-143). Supported intents: var-to-const, remove-console, add-logging. Uses the TypeScript c...

Route task to agents based on test coverage gaps (ruvector integration)

Compute attention-weighted similarity using MoE/Flash/Hyperbolic Use when native Bash hooks (via Claude Code\

Force immediate SONA learning cycle with EWC++ consolidation Use when native Bash hooks (via Claude Code\

End trajectory and trigger SONA learning with EWC++ Use when native Bash hooks (via Claude Code\

Begin SONA trajectory for reinforcement learning Use when native Bash hooks (via Claude Code\

Route task to optimal Claude model (haiku/sonnet/opus) based on complexity Use when native Bash hooks (via Claude Code\

Send cross-agent notification Use when native Bash hooks (via Claude Code\

Assess risk before executing a command Use when native Bash hooks (via Claude Code\

Record task start and get agent suggestions with intelligent model routing (ADR-026) Use when native Bash hooks (via Claude Code\

Analyze repository to bootstrap intelligence (4-step pipeline) Use when native Bash hooks (via Claude Code\

Initialize a new session and auto-start daemon Use when native Bash hooks (via Claude Code\

Agent Teams hook — fired when a teammate agent finishes its turn; reports whether a pending task can be auto-assigned. Use when native Task is wrong because you have a persisten...

Dispatch a background worker for analysis/optimization tasks Use when native Bash hooks (via Claude Code\

Embed hierarchical structure in Poincare ball. Uses hyperbolic geometry for optimal tree representation with logarithmic distortion.

Build and query entailment graphs using hyperbolic embeddings. Supports transitive closure and pruning strategies.

Advance a firmware rollout to next stage

Spin up an Anthropic-managed cloud agent (Agent + Environment + Session) — the CLOUD counterpart of wasm_agent_create. Use when wasm_agent_create (local WASM sandbox) is wrong b...

Send a user turn to a managed cloud-agent session and wait for it to go idle, returning the assistant text + a tool-use trace — the CLOUD counterpart of wasm_agent_prompt. Use w...

ADR-153 — Darwin Mode: mutate one of seven harness policy surfaces (planner/contextBuilder/reviewer/retryPolicy/toolPolicy/memoryPolicy/scorePolicy), sandbox-score each variant,...

Adversarial red/blue LLM testing via @metaharness/redblue — generates attacks across OWASP LLM Top-10 / NIST AI RMF families (prompt injection, tool misuse, data leakage, jailbr...

ADR-153 — upstream Darwin Shield (their own ADR-155): evolves a champion security-detection harness against a 10-vuln/9-decoy ground-truth corpus and grades on TPR/FPR/patch-pas...

Optimize neural model performance Use when nothing native trains on your workflow — Claude Code has no learning loop. Use to train SONA/MoE/EWC patterns from successful task out...

Make predictions using a neural model Use when nothing native trains on your workflow — Claude Code has no learning loop. Use to train SONA/MoE/EWC patterns from successful task...

Train a neural model Use when nothing native trains on your workflow — Claude Code has no learning loop. Use to train SONA/MoE/EWC patterns from successful task outcomes; query ...

Run performance benchmarks Use when native shell timing (

Apply performance optimizations Use when native shell timing (

Profile specific component or operation Use when native shell timing (

Perform causal inference using do-calculus

Compute quantum topology features (Betti numbers, persistence)

Solve combinatorial optimization using quantum annealing simulation. Supports QUBO, Ising, SAT, Max-Cut, TSP, and dependency problems.

Resolve complex dependency graphs using quantum-inspired optimization. Handles version conflicts, minimizes package size or vulnerabilities.

Grover-inspired search with quadratic speedup for unstructured search problems. Provides O(sqrt(N)) query complexity.

Optimize using Quantum Approximate Optimization Algorithm. Best for Max-Cut, portfolio optimization, scheduling, and routing problems.

Optimize task scheduling using quantum algorithms. Minimizes makespan, cost, or maximizes resource utilization with dependency constraints.

Execute attention mechanism (39 types: multi-head, flash, sparse, linear, etc.)

Execute GNN layer (GCN, GAT, GraphSAGE, GIN, MPNN, EdgeConv)

Hyperbolic embedding operations (Poincare ball, Lorentz hyperboloid)

Self-learning query optimization and index tuning

Adapt MicroLoRA weights with quality feedback. Use when sending every prompt to the Anthropic API is wrong because you need local inference — air-gapped environments, MicroLoRA-...

Run SONA instant adaptation with a quality signal. Use when sending every prompt to the Anthropic API is wrong because you need local inference — air-gapped environments, MicroL...

Build RVF container

Initialize a swarm with persistent state tracking Use when native Task tool is wrong because you need multi-agent coordination — topology (hierarchical/mesh/star), consensus (ra...

Orchestrate a task

Approve a submitted plan

Route multiple tasks to teammates optimally, avoiding over-assignment.

Enable BMSSP-powered optimization for team topology and task routing.

Launch swarm to execute an approved plan

Push team to Claude.ai remote session

Route a task to the best-suited teammate using semantic matching

Spawn a new teammate in a team. Returns AgentInput for Claude Code Task tool.

Teleport team to a new context/working directory

Create a new terminal session Use when native Bash is wrong because you need a persistent terminal session across turns/agents with output capture and replay. For one-shot shell...

Execute a command in a terminal session Use when native Bash is wrong because you need a persistent terminal session across turns/agents with output capture and replay. For one-...

ADR-175-inspired Test-Driven Repair. Use when you have a failing test and want a verified, bounded-cost fix. Spawns a bounded headless

Create a sandboxed WASM agent with virtual filesystem (no OS access). Optionally use a gallery template. Use when native Task is wrong because the workload needs sandboxed isola...

Send a prompt to a WASM agent and get a response. Use when native Task is wrong because the workload needs sandboxed isolation — untrusted code execution, browser-side run, dete...

Execute a tool on a WASM agent sandbox. Tools: read_file, write_file, edit_file, write_todos, list_files. Use flat format: {tool, path, content, ...}. Use when native Task is wr...

Create a WASM agent from a gallery template. Use when native Task is wrong because the workload needs sandboxed isolation — untrusted code execution, browser-side run, determini...

Execute a workflow Use when native TodoWrite + sequential Bash is wrong because the work has a real dependency graph that needs persistence, retry policy, pause/resume, and step...

Pause a running workflow Use when native TodoWrite + sequential Bash is wrong because the work has a real dependency graph that needs persistence, retry policy, pause/resume, an...

Resume a paused workflow Use when native TodoWrite + sequential Bash is wrong because the work has a real dependency graph that needs persistence, retry policy, pause/resume, an...

Run a workflow from a template or file Use when native TodoWrite + sequential Bash is wrong because the work has a real dependency graph that needs persistence, retry policy, pa...