// SCAN REPORT

Exocortex

44 tools. 22 can modify or destroy data without limits.

1 destructive tool with no built-in limits. Policy required.

Last updated: 30 Jun 2026

22 can modify or destroy data

22 read-only

44 tools total

Community server · catalogue entry verified 30/06/2026

How to control Exocortex ↓

TOOL MAP

What Exocortex exposes to your agents

Read (22) Write / Execute (21) Destructive / Financial (1)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Exocortex tools

Destructive · High memory_forget This tool permanently removes data (memories) from persistent storage without the ability to undo the operation. Execute · Medium memory_benchmark The tool triggers execution of a benchmark suite with measurable side effects (computation, resource consumption, potentially I/O operations against the memory system). Execute · Medium memory_compile This tool executes a procedural operation to reorganize and process stored memory data rather than simply reading or writing individual entries. Execute · Medium memory_maintenance The tool runs backend operations that modify system state (importance adjustments, archival actions) and supports conditional execution via flags.

22 of Exocortex's 44 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Exocortex

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

Deny destructive operations

{
  "memory_forget": {
    "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

{
  "agent_task": {
    "limits": [
      {
        "counter": "agent_task_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "diary_list_agents": {
    "limits": [
      {
        "counter": "diary_list_agents_per_minute",
        "window": "minute",
        "max": 60,
        "scope": "grant"
      }
    ]
  }
}

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

Create a free account and register Exocortex — nothing to install.
Add these rules — paste them, or build them visually. Tune the limits to your setup.
Point your MCP client (Claude, Cursor, anything) at your gateway URL.

ENFORCE POLICY ON EXOCORTEX →

Instant setup, no code required.

FULL CATALOGUE

All 44 Exocortex tools

DESTRUCTIVE 1 tools

Destructive memory_forget Delete a memory from Exocortex by ID. Use when information is outdated or incorrect.

EXECUTE 4 tools

Execute memory_benchmark Run a retrieval benchmark measuring recall@5, recall@10, and MRR. Execute memory_compile Compile the memory system into a browsable wiki of interlinked markdown articles. Execute memory_maintenance Run maintenance tasks: importance adjustment, archival, health checks, and optional operations via boolean fla Execute memory_security_audit Run a behavioral security audit on the memory store.

WRITE 17 tools

Write agent_task Manage the shared agent task queue. Create tasks for other agents, claim your assigned tasks, report results, Write diary_write Record a diary entry for an agent. Call after each session to log what happened, Write goal Manage goals: create, list, update, get details, log progress, and manage milestones. Write memory_clip Quick-ingest a URL into the knowledge base. Fetches the page, extracts content, chunks it, and stores with imm Write memory_contradictions List or resolve detected contradictions between memories. Contradictions are detected nightly and can be dismi Write memory_correct Correct a memory that contains wrong or outdated information. Creates a new memory with the corrected content, Write memory_feedback Record explicit retrieval-quality feedback: a specific memory was perfectly relevant / wrongly retrieved / out Write memory_ingest Ingest content into Exocortex from files, URLs, or session transcripts. Write memory_job_health Record job execution outcomes and query job health. Use record_outcome to log a job run, or omit it to get a h Write memory_link Create or remove a link between two memories. Links enable graph-aware context retrieval — linked memories sur Write memory_obsidian_sync Incrementally sync memories to an Obsidian vault. Only writes files for memories created or updated since the Write memory_promote Promote a synthesis, analysis, or set of memories into a persistent wiki article. Use when a query result, com Write memory_store Store a new memory. Set provider/model_id/agent for attribution, namespace for project scope, deduplicate: tru Write memory_update Update an existing memory Write prediction Manage predictions: create forecasts, list/filter, resolve outcomes, get details, or view calibration stats. Write wiki_append_article Append or replace a dated section in a journal-style wiki article. Write wiki_write_article Write or update a synthesized wiki article. Use after reading an extractive article

READ 22 tools

Read diary_list_agents List all agents that have diary entries, with entry counts and last entry date. Read diary_read Read an agent Read embedding_models List available embedding models with their characteristics. Read memory_auto All-in-one memory tool. Understands natural language: Read memory_browse Browse memories without semantic search. Filter by tags, content type, or date range. Returns most recent firs Read memory_context Load broad context about a topic for session start or subject switching. Returns memories ranked by relevance Read memory_diff See what changed since a timestamp — new, updated, and archived memories. Read memory_entities List tracked entities (people, projects, technologies) with linked memory counts. Read memory_facts Query structured subject-predicate-object facts. Use for precise lookups: ports, versions, config values, depe Read memory_get Fetch full content for specific memory IDs. Implicitly signals usefulness for future ranking. Read memory_graph Analyze the entity relationship graph — centrality metrics, bridge entities, knowledge domain connections. Read memory_lint Run a comprehensive knowledge-base health check. Returns contradictions, stale claims, orphan entities, unlink Read memory_navigate Browse the memory palace — hierarchical view of all knowledge organized into wings (projects), Read memory_ping Health check — returns memory counts, entity/tag stats, and uptime. Read memory_project_snapshot Get a quick project snapshot: recent activity, active goals, recent decisions, open threads, and learned techn Read memory_reason Assemble a structured reasoning brief from retrieved memories. Returns evidence + a synthesis rubric — the CAL Read memory_research Search the web for a topic and ingest best sources as chunked reference knowledge. Read memory_search Hybrid search (semantic + keyword + graph). Use compact: true for previews, expanded_query for better recall, Read memory_timeline Query decision history, memory lineage, topic evolution, or temporal hierarchy. Read memory_wakeup Load compressed wake-up context (~200 tokens). Returns identity, active projects, Read memory_wiki_refresh Refresh wiki articles that have stale content (memories updated since last compile). Only recompiles affected Read unified_search Search across all retrievable surfaces in the exocortex DB at once: memories, predictions, contradictions, goa

EXPLORE

Related servers

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

FAQ

Questions about Exocortex

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

Yes. The Exocortex server exposes 1 destructive tools including memory_forget. 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 Exocortex? +

The Exocortex server has 17 write tools including agent_task, diary_write, goal. 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 Exocortex.

How many tools does the Exocortex MCP server expose? +

44 tools across 4 categories: Destructive, Execute, Read, Write. 22 are read-only. 22 can modify, create, or delete data.

How do I enforce a policy on Exocortex? +

Register the Exocortex 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 Exocortex tool call.

Deterministic rules across all 44 Exocortex tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON EXOCORTEX →

Instant setup, no code required.

44 Exocortex tools catalogued and risk-classified — across an index of 43,000+ MCP servers.