tandem_recommend

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

With no annotations, the description carries the full load and does so: preview is the default with 'exact bounded payload without a key or provider call', while sending requires enablement, an OpenRouter key, preview=false and a matching expected_input_sha256 from the approved preview. It flags the important limits directly — 'The hash binds input; it does not establish human approval', 'Supplied goal and candidate descriptions may contain secrets; this is not redaction', 'Exact duplicates reuse the attempt; uncertainty never authorizes replay' — plus journal retention of request/result and the distinction between candidate/none/unclear and provider failure.

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?

Purpose, default behavior, send gating and safety caveats are front-loaded in that order, and nearly every sentence introduces a distinct constraint rather than restating the schema. The telegraphic, one-clause-per-sentence style is dense and would scan faster as a short bulleted list, but there is no filler to cut.

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 nested-schema tool with zero annotation and zero schema-description coverage, this is close to complete: it covers default vs send paths, credential and enablement prerequisites, hash provenance, side effects (journal retention), duplicate handling and failure taxonomy, and an output schema exists so return shape need not be restated. The remaining gap is the 'kind' parameter's two modes and any guidance on candidate count/id conventions, which the caller must infer.

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%, so the description must compensate, and it does for the two highest-risk parameters: preview's default/bounded behavior and expected_input_sha256's required matching provenance along with the warning that the hash is not approval. It also notes that goal and candidate descriptions carry potentially sensitive free text. It leaves 'kind' (skill vs review_direction) unexplained beyond the first sentence's phrasing and gives no size bounds, so it falls just short of complete.

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 first sentence states a specific verb and resource: 'Suggest one skill or review direction from an explicit small candidate list', which also bounds the operation to a caller-supplied list rather than open discovery. That scope ('explicit small candidate list', 'No skill discovery') distinguishes it from the discovery/context siblings in the namespace. An agent can decide between this and tandem_findings or tandem_scope without opening the 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?

It states the actor and timing ('Coordinator-only advisory operation, available before any task exists') and then enumerates what the tool does not do: 'No skill discovery, file/context/history reads, launches or permission changes.' It also names an explicit exclusion ('Do not inject this advice into an independent snapshot-review stage'), so both when-to-use and when-not-to-use are covered.

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