start_run

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

With no annotations, the description carries the full burden and delivers: async lifecycle, immediate success/failure signaling (a config error or a path outside NETDRYRUN_MCP_ROOT is answered at once with status error and no run_id), statefulness ('the run lives in this server: if the server restarts, start again'), expected duration (60–120 s), and hard environmental prerequisites (Linux/WSL2, Docker, containerlab, root-equivalent rights; fails on macOS and non-DinD sandboxes). It also discloses the destructive tail (the lab is destroyed at the end unless the same keep_lab semantics as dry_run apply).

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 the core identification and the immediate return contract, followed by polling rules, failure semantics, and prerequisites in a sensible order. Some material is server-general rather than tool-specific ('Ping is not a test', 'Never invent or weaken checks... call record_baseline'), which slightly dilutes focus, but every clause is substantive.

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 the description needn't document return values, yet it still explains the immediate running/error shapes. Polling, retry limits, restart loss, timing and platform prerequisites are all covered. The remaining gap is the two input parameters' meaning, especially keep_lab, which is invisible in both schema and description.

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% for both parameters, and the description only says 'Same `path` and keep_lab as dry_run' — deferring rather than explaining. It never says what path points to (config file? directory?) or what keep_lab actually does (leaving a deployed lab behind has cost and cleanup implications). The deferral is a small mitigation, but an agent must go read dry_run to fill the gap.

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 a precise framing — 'The asynchronous dry_run, for clients that give up on a long tool call' — that names the verb (a dry run), the mode (async), and the exact sibling it mirrors. It states the immediate return shape ('returns at once with a `run_id` and status running (no verdict yet)') and that the same deploy/check/destroy lifecycle follows, so an agent can distinguish this from dry_run and get_result without opening a 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?

Routes explicitly: call get_result with the run_id every poll_after_s until status is no longer running, and use this instead of dry_run when the client gives up on long calls. Also gives decision rules for outcomes — INCONCLUSIVE means stop, retry at most once, and error/inconclusive is a setup problem unrelated to the change — plus the server-level trigger ('before proposing or applying any routing or policy change') and an anti-pattern ('Ping is not a test').

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