mnema

Official

Overview Schema Source Related Servers Discussions Admin

Record a decision

record_decision

Record an immutable decision with its rationale and rejected alternatives, so later readers can understand why a choice was made and search whether an option was already declined.

Instructions

Record a decision into the mnema chain, attributed to this agent and pinned to the current session. A decision needs both a title and a rationale (why it was made). Optionally record what you considered and turned down, and why not (alternatives): the reasoning a later reader needs when somebody proposes the thing you already rejected; it is searchable, so search answers "did we already turn this down?". A decision is immutable, so it is recorded now: an option rejected later is a new decision, or supersedes this one. Optionally pick the scope it lands in: public (team-visible), private (this machine, this project), or global (personal, cross-project); omitted, a decision is a declaration about the project and lands PUBLIC: committed, so the team gets it on clone (global outside a project); it is born proposed, so recording one proposes it rather than settling it. Optionally pick the project it lands in, when the workspace holds more than one. Returns the citable ADR- label (a decision has no short alias) and which tree it landed in. RECORDING IS PERMANENT: mnema is append-only and nothing deletes a fact. A record in the public scope is committed to the repository and reaches every machine that clones it (published, if that repository is public). So do NOT record credentials here. Values in a recognized format (cloud keys, API tokens, private keys, a password inside a URL) are replaced with a typed placeholder before anything is written and the reply says what was replaced, but a format mnema does not recognize is written verbatim and cannot be taken back. Do not record personal data either: an email address is replaced with and reported, and the mnid already says who wrote. Each text field holds at most 65536 bytes; a longer one is refused, not truncated.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "title": "<title>",
      "rationale": "<rationale>"
    },
    "name": "record_decision"
  }
}

Input Schema

Name	Required	Description
`scope`	No	Where the decision lands; overrides the routing rule (public).
`title`	Yes	The decision title.
`project`	No	Which project of this workspace the fact belongs to: its absolute path, as this server gave it; a bare name is refused. Omitted, it lands in the project this session resolved to (the one `bootstrap` names). Pass it whenever the work being recorded happened in a different project of this workspace.
`rationale`	Yes	Why the decision was made.
`alternatives`	No	What was considered and turned down, and WHY NOT: both halves; the name of a rejected option without its reason loses most of the value. Omit it when there was no real alternative.

Schema Changelog

Changes observed during successful MCP inspections.

First observedNov 9, 2026

TDQS

A4.2/5.0

Behavior3/5

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

Annotations already declare a non-read-only, non-idempotent, non-destructive write, and the description is consistent with all of them while adding real context: append-only permanence, public-scope commits reaching every clone, and birth state `proposed`. It also documents secret/email redaction behavior and the 65536-byte refusal, which the annotations cannot express. However, the 'nothing deletes a fact' permanence statement and the redaction caveats are the substantive additions; the middle prose largely restates scope routing that the schema already carries.

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

Conciseness4/5

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

Purpose is front-loaded and the most decision-relevant warnings (permanence, no credentials) are stated emphatically. It is still a single dense block of very long sentences, with some repetition of the public-scope consequence ('committed, so the team gets it on clone' vs. 'committed to the repository and reaches every machine that clones it'), which costs a little economy.

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

Completeness5/5

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

With no output schema, the description still covers the return value (the citable ADR-<n> label and which tree it landed in), plus the redaction report in the reply. For an irreversible, repo-publishing write it supplies the prerequisites, defaults, failure mode (over-length refused, not truncated), and security exclusions an agent needs to call it safely.

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

Parameters4/5

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

Schema coverage is 100%, so the schema already documents all five parameters and the baseline is 3. The description goes beyond it by explaining the consequence of omitting `scope` (defaults to public, committed, team-visible on clone, global outside a project), why `alternatives` needs both halves to be searchable, and when to pass `project` (a workspace with more than one project). That is meaning the schema text does not supply.

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?

The first sentence names a specific verb and resource ('Record a decision into the mnema chain') and adds two scoping facts (attributed to this agent, pinned to the current session). It also distinguishes itself from the `search` sibling by explaining that `alternatives` is what makes the 'did we already turn this down?' query answerable, so an agent can route without opening a 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?

Usage context is clear: record when a choice was made, include `alternatives` when a real option was rejected, omit it otherwise, and treat a later rejection as 'a new decision, or supersedes this one'. The only gap is that it never explicitly contrasts recording with `decision_transition` for the settle/propose lifecycle, leaving that inference to the reader.

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