Skip to main content
Glama
navig-run

NAVIG

Official
by navig-run
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Local

Recall: past work and decisions

navig_recall
Read-onlyIdempotent

Search recorded past conversations, decisions, plans, ideas, tables, and open questions, with every result cited. Retrieve record details, related items, timelines, or source excerpts by id.

Instructions

Search the owner's recorded past: conversations, decisions, tables, open questions, forgotten ideas and source excerpts, every result cited. Read-only. Use navig_knowledge for docs/wiki and navig_memory for durable facts. Returns records with ids and citations; pass an id to get/related/timeline or a citation to source_excerpt.

Input Schema

NameRequiredDescriptionDefault
idNoRecord id (rk-…), from any other recall tool. Used by: get, related, timeline.
rawNoSearch conversation turns, not records. Used by: search.
typeNoOnly this record type (decision, plan, idea, table, finding, feature, question, …). Used by: search.
afterNoConversation turns to include after the cited message (default 3) Used by: source_excerpt.
limitNoMaximum results (1-50). Used by: search, ask, decisions, tables, open_questions, forgotten_ideas.
queryNoWords to find; "quoted text" matches a phrase. Used by: search.
actionYesOperation to perform. Required parameters per action: search: query; get: id; ask: question; related: id; timeline: id; decisions: none; tables: none; source_excerpt: source_id, msg_id; open_questions: none; forgotten_ideas: none. What each action does: search: Search the owner's past AI conversations (Claude Code, Codex, shared chats, session audits) for decisions, plans, ideas, tables and findings; falls back to the raw conversation tur (required: query) | get: One record in full: summary, body, every piece of evidence (verbatim, cited), its timeline and related records (required: id) | ask: Has the owner asked this before? Returns earlier questions the owner typed, the reply that followed each one, and matching decisions (required: question) | related: Records linked to one record (evolves, supersedes, related, contradicts, answers) (required: id) | timeline: The append-only history of one record: when it was created, merged, endorsed or superseded, and by which kind of evidence (required: id) | decisions: Decisions, newest first: the owner's answers to structured questions and audited decisions (required: none) | tables: Markdown tables from past conversations, verbatim (the body is the table) (required: none) | source_excerpt: The original conversation around one cited message (default 3 turns either side) — check a citation before relying on it (required: source_id, msg_id) | open_questions: Questions still unresolved, and recorded conflicts nobody settled (required: none) | forgotten_ideas: Ideas nobody has shown were built (impl not_found or unknown) and that were not rejected, the longest-untouched first (required: none)
beforeNoConversation turns to include before the cited message (default 3) Used by: source_excerpt.
msg_idNoMessage id from the same citation. Used by: source_excerpt.
statusNoOnly this status (approved, confirmed, rejected, …). Used by: search, decisions.
projectNoProject id to read instead of the one the client is working in (shared is always included). Used by: search, ask, decisions, tables, open_questions, forgotten_ideas.
questionNoThe question, in plain words. Used by: ask.
source_idNoSource id from a citation (e.g. cc:<session>). Used by: source_excerpt.
all_projectsNoRead every project, not just the current one plus shared. Ask for it explicitly. Used by: search, ask, decisions, tables, open_questions, forgotten_ideas.
evidence_limitNoEvidence items to return. Used by: get.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Addedv0.1.3

TDQS

A4.2/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already cover the safety profile (readOnlyHint, idempotentHint, destructiveHint=false), so the description only needs to add context. It does: results are always citation-backed, records are returned with ids, and those ids/citations chain into get/related/timeline/source_excerpt. It does not mention rate limits, project-scoping effects, or result caps beyond what the schema states.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences with no filler: purpose first, safety and routing next, return-and-chain behavior last. Every sentence carries information the agent needs, and the verb leads the description.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 15-parameter, 10-action tool with no output schema, the description is nearly sufficient: it states the result shape (records with ids and citations) and the follow-up path. Deeper per-action detail lives in the enum, which is acceptable, though a note on search-vs-raw fallback behavior would close the remaining gap.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and the schema even maps required params per action, so the description cannot add much. Its one contribution ('pass an id to get/related/timeline or a citation to source_excerpt') restates the schema's own 'Used by' annotations rather than adding new meaning, which is the baseline 3 case.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

Front-loads a specific verb (Search) and resource (the owner's recorded past), then enumerates the content classes it covers (conversations, decisions, tables, open questions, forgotten ideas, source excerpts). It explicitly differentiates from sibling tools by naming navig_knowledge and navig_memory, so an agent can select it without opening any schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Names two alternatives and the conditions that select them ('navig_knowledge for docs/wiki and navig_memory for durable facts'), which is strong alternative routing. It stops short of a full when-not statement (e.g., no guidance on when to fall back to raw browsing), so it is clear context rather than complete exclusions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Deploy Server

Other Tools