Ricursor
Server Details
Long-term, rotatable project memory and pre-write checks for AI coding agents.
- Status
- Healthy
- Uptime
- 100.0% over the last 24 hours
- Last Tested
- Transport
- Streamable HTTP · MCP 2025-11-25
- URL
- ChatGPT
- –
- Claude
- –
TDQS
Scored across 12 tools
The action/group action pairs are well delineated by their check/consume split, but ricursor_check vs ricursor_action_check are easy to confuse, as are the read-only metadata tools ricursor_context, ricursor_ledger, and ricursor_snapshot. Descriptions do disambiguate on close reading, but the names alone do not make the boundaries obvious.
Every tool carries the same ricursor_ prefix and snake_case formatting, and the action_check/action_consume/action_group_check/action_group_consume family is highly predictable. The minor deviation is that half the tools are bare nouns (context, ledger, handoff, report, snapshot, workspaces) while others use verb-led names, so the verb_noun pattern is not uniform.
Twelve tools is well within a reasonable range and each maps to a phase of the broker lifecycle (snapshot, check, consume, correct, report, handoff). A few record-keeping tools (handoff, report, ledger) feel adjacent enough that the surface could be slightly leaner, but nothing is clearly redundant.
The core gate lifecycle is covered end to end: metadata sync, single and group checks, atomic receipt consumption, corrections after a block, context/ledger reads, and usage/state reporting. Gaps are minor, such as no explicit way to inspect or expire outstanding receipts or to query block history beyond the ledger.
Available Tools
12 toolsricursor_action_checkPropose actionBInspect
Metadata-only decision for a JS/TS write or delete. Returns allow or block (with reasons to fix) plus a one-use receipt. Never send source.
Inputs: kind (required), path (required), project (required), action_id (required), +5 more.
ricursor_action_consumeConsume receiptAInspect
Atomically consumes a receipt immediately before the broker executes. Fails if anything changed.
Inputs: receipt (required), action_id (required), result_hash (required), baseline_hash (required).
ricursor_action_group_checkPropose related changesCInspect
Metadata-only decision for up to 20 related JS/TS writes/deletes judged on their combined end state (moves, renames).
Inputs: actions (required), project (required), group_id (required).
ricursor_action_group_consumeConsume group receiptCInspect
All-or-nothing consumption before the broker writes the group.
Inputs: receipt (required), group_id (required).
ricursor_checkCheck planned writeCRead-onlyInspect
Stop on block or errors. Checks tracked versions, exports and accepted import boundaries; never writes code.
Inputs: hash (required), path (required), project (required), metadata (required), +2 more.
ricursor_contextServe contextBRead-onlyInspect
The verified corrections and recent file changes an agent must respect before it writes code.
Inputs: task, limit, since, project, +1 more.
ricursor_correctSave verified correctionAInspect
After Ricursor blocks a check, save the one-sentence lesson so every agent on this project gets it. Only accepted for your own blocked action from the last 24 hours.
Inputs: rule (required), project (required), action_id (required).
ricursor_handoffSave model handoffCInspect
Save structured task IDs and versioned rule decisions; never include prompts or source.
Inputs: state (required), project (required), task_id (required), expected_version (required).
ricursor_ledgerRead model alignmentBRead-onlyInspect
Read scoped metadata and rule versions before switching models. Observe only.
Inputs: project (required).
ricursor_reportReport activityCInspect
Record what an agent spent (usage) and what it changed (state).
Inputs: request_id (required), state, usage, project, +1 more.
ricursor_snapshotSync repositoryCInspect
Sync locally extracted JS/TS structural metadata only; never transmit source.
Inputs: files (required), project (required).
ricursor_workspacesList workspacesCRead-onlyInspect
The workspace this key can reach.
Inputs: no named parameters.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Related MCP Connectors
Project memory, documentation and controlled secret access for AI coding agents.
Project memory for coding agents: requirements, decisions, code graph and delivery telemetry.
Evidence-backed project memory and checkpoints for AI coding agents.
Shared, scoped, reviewable team memory for AI coding agents.
Related MCP Servers
- AlicenseAqualityDmaintenanceLocal-first project memory for AI coding agents. Records failed attempts, fragile files, and decisions per repo, and warns the agent via hooks before it repeats a recorded mistake.669 npmMIT
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to carry durable project memory across windows, agents, and days, injecting a compressed context of architecture, progress, decisions, and critical files so a new session can resume without re-reading the whole repository. It also classifies files as current/legacy/intermediate and suggests cleanup so stale or draft artifacts don't mislead the agent.8 npmMIT
- AlicenseNot gradedqualityBmaintenanceProvides persistent memory for coding agents and grounds their claims by verifying against the actual codebase, preventing hallucinated responses.506 npmApache 2.0
- AlicenseNot gradedqualityBmaintenancePersistent memory for AI coding agents that knows when it has gone stale.3MIT
Glama MCP Gateway
Add one secure layer between your agents and this server.