exp_run

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

Annotations only declare destructive=true and readOnly=false; the description adds substantial behavioral context beyond that: factor values are read first and restored on abort (read back and checked), every write is re-checked against a server guard, pinned to the starting machine, and audited with symbol/value, and a refusal/target change/timed-out write aborts the experiment. It also discloses the control-class confirmation flow (elicitation/confirm_token) and the single-experiment concurrency limit. This is exactly the extra context annotations cannot carry.

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?

Information-dense but structured poorly: the opening sentence is a long arrow-delimited pipeline that is hard to parse and not obviously the highest-value content for selection. The behavioral safety details are valuable but buried mid-paragraph. Every sentence does carry meaning, so it is not padded, just densely and unfront-loadedly written.

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-value explanation is not needed. For a complex, destructive background runner the description covers the pipeline, safety/audit behavior, abort conditions, dry_run semantics, and confirmation flow. The main gap is undocumented parameters (resume, max_runs, name), which slightly under-serves correct invocation.

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% across 5 parameters, so the description must carry the burden. It explains dry_run well (resolves symbols, coerces enum names to ints, prints plan without writing) and clarifies confirm_token's role in guarded mode. But name, resume, and max_runs get no explanation at all, leaving three of five parameters undocumented.

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 opens with a clear verb+resource ('Run an experiment in the background') and then specifies the exact execution pipeline (preconditions → factor values → snapshot → stimulus → ... → runs.jsonl). This distinguishes it from siblings like exp_validate, exp_status, exp_runs and exp_abort, which clearly handle other phases. The core purpose is unambiguous, though the pipeline detail crowds the framing.

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?

Usage is implied through the described dry_run resolution/planning mode and the note that 'Only one experiment may run at a time,' plus the guarded-vs-full mode confirmation behavior. However, it never explicitly states when to prefer this over exp_validate or exp_save, nor routes the agent to exp_abort/exp_status for follow-up. Adequate but not directive.

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