propose_memory

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

With no annotations present, the description carries the full burden and it delivers: it discloses auto-approval default, the PM_REVIEW=1 escape hatch, that redaction/collision-safety are NOT relaxed, that secrets are stripped because the store is committed to git, that id clashes suffix rather than overwrite, and that auto-approved entries carry verified_by='auto'. It also warns against credentials — comprehensive side-effect disclosure well beyond the minimum.

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

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

Four dense paragraphs, front-loaded with the action and the one flag (PM_REVIEW=1) an agent must know first; the rationale sentence explains why auto-approve is the default, which prevents the agent from second-guessing the contract. Minor redundancy exists ('committed to git' appears twice) and the middle rationale paragraph could be tightened, but no sentence is wasted.

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

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

For a 10-parameter write tool with no annotations and no output schema, the description covers the essential ground: purpose, approval behavior, safety constraints, usage recipes, and the required refs/content formats. The gaps are the unexplained parameters (symbol, source, file, repository, tags) and the absence of any statement about return values or success signals, though the behavioral contract is unusually complete.

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

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

Schema coverage is 0%, so the description must compensate — and it does for the high-stakes parameters: type is given concrete values ('code_symbol', 'worklog'), refs is explained as PRD/spec/ticket references, and content gets a full WHAT/WHY/NEED/WHY THIS WAY/GOTCHA template. However, symbol, source, file, repository, and tags get no semantic explanation, leaving ambiguity around the symbol/file/repository distinction.

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

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

Opens with 'Record a new memory' — a specific verb plus resource — and immediately extends into the behavioral contract (auto-approved by default, PM_REVIEW=1 queue). All eight siblings are read/query tools (get_, search_, read_, why_), so the write verb clearly separates this tool; the concrete use cases (code_symbol per function, worklog entry, 'why does X work this way?' answers) make the purpose unmistakable.

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

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

The third paragraph is a direct usage instruction: 'Use at task completion' with a precise recipe — one type='code_symbol' entry per changed function, the WHAT/WHY/NEED/WHY THIS WAY/GOTCHA body format, refs pointing at PRD/spec/ticket, plus a worklog entry — and a second trigger ('record the answer when the user asks why does X work this way?'). It doesn't name read-sibling alternatives explicitly, but the write-vs-read split plus the concrete triggers make when-to-use unambiguous.

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