agy_discussion_extend

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

Annotations only declare it is a non-read-only, non-destructive, closed-world mutation; the description adds substantial context beyond that: it preserves spent attempts, roles, routes, memory and plan versions, records before/after limits and a reason, and states that failures still burn budget. The optimistic-concurrency signal ('expected version', 'pending calls block changes') is disclosed. It still does not say what the response contains or what happens on a version mismatch.

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?

The most decision-relevant facts (verb, per-remote cap) are front-loaded, which is good. But the prose is a semicolon-chained wall of domain shorthand that mixes constraints, side effects, and exclusions without structure, making it slower to scan than its length warrants.

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 bounded mutation with no output schema, the description covers limits, concurrency, preserved state and explicit non-effects — enough for an agent to call it safely. The main gap is that nothing is said about the returned/updated budget state or failure modes, which an agent would want after a version conflict.

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?

With 0% schema coverage the description carries the full burden, and it does most of it: the `additions` map is bounded to 1..8 per remote with a 32 ceiling, its keys are explained (gemini/claude/codex councils, 'reviewer' for peer tasks), and `expected_budget_version` is given an initial value. `discussion_id` and the relationship between `reason` and the recorded audit entry are only implied.

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?

The opening sentence states a specific verb and effect ('Explicitly add 1..8 attempts to selected task remotes') plus a hard cap, which is far more precise than the tool name alone. It also differentiates itself from execution-oriented siblings by saying it 'Does not call/retry models ... or start a loop'. The jargon ('task remotes', 'council keys', 'reviewer') is never defined, so an agent must already know the domain.

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?

'Read first; expected version is 0 initially' and 'Pending calls block changes' give concrete preconditions and sequencing. The negative clauses ('does not reopen archived tasks or start a loop', 'failure/cancellation still consumes an attempt') function as when-not guidance. It stops short of naming a sibling alternative to use instead when budget is exhausted.

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