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

Annotations only carry the safety profile (readOnly, idempotent, non-destructive); the description goes well beyond them by disclosing that the call does NOT clear the dead-man clock, does NOT move `last_event_at`, and does NOT clear the miss counter, plus how duration is measured and when `reason: "too_long"` fires. That is exactly the non-obvious behavior an agent needs.

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 and dense with real information, but it repeats itself: the dead-man-clock point is made twice ("does not clear the dead-man clock: a routine that starts and hangs still becomes a miss" and again "does not clear the miss counter, so a routine that starts and hangs still becomes a miss"). One of those sentences should be 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 single-parameter signal with no output schema, the description covers the lifecycle semantics, the alerting edge case, method support (GET or POST), and the `rid` correlation requirement. It never says what a successful call returns or what an error looks like, which is a minor remaining gap.

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 coverage is 100% and already documents `id`, so baseline is 3. The description adds meaning the schema lacks: the `?rid=` correlation parameter needed when runs overlap, and the consequence of omitting it (duration measured against the wrong start).

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 description clearly states the tool's action ("Says the run STARTED") and its role in the run lifecycle — opening a run that a later success closes. It distinguishes start from success/failure semantically, but never names the sibling signals (signal_fail, signal_exit_code, signal_slug) it should be chosen over, so an agent gets no explicit routing.

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?

Gives concrete conditions for correct use: pass `?rid=` on both ends when runs can overlap, and expect a `too_long` alert when `max_duration_sec` is exceeded. It also implies when NOT to rely on it ("start is not proof of life"), but never names alternative tools or states an explicit exclusion against siblings.

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