aiworkhub_review_packet

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

With readOnlyHint already declared, the description still adds real behavioral context: output is byte-bounded (default and p50/p90/max figures, zero truncation in the measured population), it is a snapshot of persisted facts and 'nothing here is a verdict', accepting still re-runs the fold via `accept_review`, and `accept_preview` is the one live computation and is inlined because `completion_inbox` lacks launch authority. That is meaningful disclosure beyond annotations, though the volume/usage statistics lean toward sales copy.

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?

Front-loaded with READ-ONLY and the core purpose, which is good, but the middle is padded with call-count and Bash-re-derivation breakdowns that inform tooling economics more than tool use. The 'nothing here is a verdict' and accept_preview sentences do earn their place; the statistical sprawl does not.

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?

An output schema exists, so return values need no prose, and the description does enumerate the six nested objects and state the byte bound. The gap is input-side: for a tool whose only required argument is undocumented, an agent cannot know what `request_id` must be or how to size `max_bytes`, which is not covered elsewhere.

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 description coverage is 0% and none of the three parameters are explained in the description. The required `request_id` is never mentioned at all, `task_id` is only implied by 'candidate', and `max_bytes` is only alluded to via 'under a measured byte bound' with no statement of what happens on truncation or how the 90000 default relates to the reported max of 13,799 B.

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?

States a specific verb and resource — returns ONE bounded review packet for a `review_ready` candidate — and enumerates exactly what that packet contains (validation rows, gate checks/blockers, MCP gate verdict, required outputs, diff checks, tier/per-lens status, deduped findings, reachability). An agent can distinguish this from `completion_inbox`, `task_review_summarize`, or `task_review_queue` 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.

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

Clear context: it is the manager's consolidated read for deciding a candidate, and it explicitly says it replaces 10 task-scoped calls (p50) per candidate plus 1,688 session-wide Bash re-derivations, which tells the agent when it beats ad-hoc investigation. It also notes the `completion_inbox` boundary. It stops short of naming the specific alternative tools it supersedes (e.g. task-scoped reads, `task_review_summarize`) or stating exclusions.

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