// SCAN REPORT

Lore Context

34 tools. 12 can modify or destroy data without limits.

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

Last updated: 29 Jun 2026

12 can modify or destroy data

22 read-only

34 tools total

Community server · catalogue entry verified 29/06/2026

How to control Lore Context ↓

TOOL MAP

What Lore Context exposes to your agents

Read (22) Write / Execute (10) Destructive / Financial (2)

Running more than one server? Check your whole stack →

MOST DANGEROUS TOOLS

Critical Risk

The most dangerous Lore Context tools

Destructive · High memory_forget While the tool permits soft-deletion (which may be technically reversible depending on implementation), the primary function is to remove data by explicit selector, matching the Destructive category definition. Destructive · High memory.delete This tool irreversibly removes data (memories) from an AI agent's context store. Execute · High eval_run This tool executes an evaluation workflow against uploaded sessions or recipe datasets. Write · High memory_supersede This tool creates a new memory version and modifies state (marking previous as superseded), which is a reversible Write operation.

12 of Lore Context's 34 tools can modify, destroy, or commit something on every call — and an agent calls them with no built-in limits.

POLICY

How to control Lore Context

PolicyLayer is an MCP gateway — it sits between your AI agents and Lore Context, 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

{
  "memory_supersede": {
    "limits": [
      {
        "counter": "memory_supersede_per_hour",
        "window": "hour",
        "max": 30,
        "scope": "grant"
      }
    ]
  }
}

Prevents bulk unintended modifications from agents caught in loops.

Cap read operations

{
  "code.callgraph": {
    "limits": [
      {
        "counter": "code.callgraph_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 Lore Context — 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 LORE CONTEXT →

Instant setup, no code required.

FULL CATALOGUE

All 34 Lore Context tools

DESTRUCTIVE 2 tools

Destructive memory_forget Forget memories by explicit selector. Requires a reason and defaults to soft delete. Destructive memory.delete Delete memories by explicit selector. Requires a reason and defaults to soft delete.

EXECUTE 1 tools

Execute eval_run Run retrieval evaluation over uploaded sessions or a v12/v13 recipe dataset.

WRITE 9 tools

Write memory_supersede Create a new version of a memory and mark the previous one superseded. Write memory_update Edit an existing memory. Risky content returns to the governance review queue. Write memory_write Save a user-approved or project-relevant memory with audit logging. Write memory.add_candidate Add a beta memory candidate through governed memory writeback. Write memory.inbox_approve Approve a memory candidate from the governance review queue. Write memory.inbox_reject Reject a memory candidate from the governance review queue. Requires a reason. Write profile.update_candidate Submit a beta profile update candidate when the profile API is available. Write source.pause Pause a hosted capture source by writing a paused heartbeat. Write source.resume Resume a hosted capture source by writing an active heartbeat.

READ 22 tools

Read code.callgraph BFS call-chain subgraph from a given symbol up to the requested depth (1-5). Returns nodes and edges. Read code.context Given a fully-qualified symbol name, returns the symbol metadata plus all incoming references (callers/importe Read code.impact Blast-radius analysis: transitive closure of code.references up to depth N. Hard cap of 500 symbols with pagin Read code.list_repos List all code repos indexed in the current vault. The last_indexed_at field is the staleness signal — compute Read code.references Reverse-index lookup: all symbols in the vault that reference the given target symbol. Same shape as incoming[ Read code.search Hybrid BM25 + dense-embedding search over all indexed code symbols in the current vault. Returns top-K matches Read code.search_text Regex/grep search over symbol name and doc_comment columns only (NOT function bodies — bodies are never stored Read context_query Compose governed, agent-ready context for a user query from memory, web, repo evidence, and tool traces. Use t Read evidence.trace_get Read an Evidence Ledger for a context trace id. Read memory_export Export visible memories as Lore JSON or Markdown. Read memory_get Read one visible Lore memory by id. Read memory_list List visible memories with pagination and filters. Read memory_search Search visible Lore memories directly. Read memory.inbox_list List memory candidates currently waiting for governance review. Read memory.recall Hosted MCP recall entrypoint. Composes governed, agent-ready context and returns its trace id for audit. Read memory.suggest Given a user message in context, returns the top-K canonical memories the host should consider injecting into Read profile.get Read the beta profile contract when the profile API is available. Read repo.detect_changes Map a unified diff to indexed files/symbols and return affected upstream symbols and suggested tests from Lore Read repo.graph_query Query Lore Read repo.impact Analyze blast radius for an indexed symbol or file using reverse repo graph traversal. Returns callers, upstre Read repo.symbol_context Return a symbol definition with incoming callers/references, outgoing callees/imports, related tests, and proc Read trace_get Read a context or memory trace by id.

EXPLORE

Related servers

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

FAQ

Questions about Lore Context

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

Yes. The Lore Context server exposes 2 destructive tools including memory_forget, memory.delete. 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 Lore Context? +

The Lore Context server has 9 write tools including memory_supersede, memory_update, memory_write. 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 Lore Context.

How many tools does the Lore Context MCP server expose? +

34 tools across 4 categories: Destructive, Execute, Read, Write. 22 are read-only. 12 can modify, create, or delete data.

How do I enforce a policy on Lore Context? +

Register the Lore Context 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 Lore Context tool call.

Deterministic rules across all 34 Lore Context tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.

ENFORCE POLICY ON LORE CONTEXT →

Instant setup, no code required.

34 Lore Context tools catalogued and risk-classified — across an index of 43,000+ MCP servers.