compute.node_selftest

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

Beyond the lone readOnlyHint=false annotation, the description discloses the duration (~10s), the concrete side effect (creates and removes one temp dir), and a firm negative guarantee (never submits a job). It still doesn't say what a failure looks like or whether partial failures abort the run, so it stops short of 5.

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?

A single dense sentence, front-loaded with the payoff (prove the control path in ~10s) and followed by the checklist and the safety guarantee. Every clause carries information; nothing is padding.

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?

There is no output schema, so the description is the only place a result shape could be described — it says which subsystems are exercised but not how results are reported (per-check pass/fail, error text). Combined with the missing node_id guidance, it is adequate but incomplete for a diagnostic tool.

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?

The single parameter node_id has 0% schema description coverage and the description never mentions it — not its format, nor that it can be obtained from compute_list_nodes. With only one obvious parameter the gap is small in practice, but the description contributes no meaning here.

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?

"Prove the control path to a node" gives a specific verb (prove/verify) and resource (node control path), and the follow-on list (ssh, POSIX shell, work_root write/read/cleanup, python3, GPU) makes the scope concrete. It does not, however, distinguish itself from close siblings such as compute_inspect_node, compute_doctor, or compute_verify_environment.

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?

"Prove the control path... in ~10s" implies a cheap pre-flight check, and "never submits a job" implies the line between this and compute_submit_run. But there is no explicit when-to-use statement and no named alternative for the many similar diagnostic siblings.

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