list_experiment_runs

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

With no annotations, the description carries the behavioral burden and does so well: it discloses ordering ('Newest first'), that grids are omitted from results, what status actually means ('whether the fan-out finished, not whether every cell succeeded'), the meaning of window and resolved_sessions, and a hard cap of 200 with no further paging and possible truncation. It does not cover authorization or permission requirements, which is the remaining gap.

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 is front-loaded, but the body is dense, line-wrapped prose mixing domain definitions (what a scoring run is) with parameter documentation, followed by Sphinx-style :param: blocks and a boilerplate object-store paragraph. Several sentences, such as the explanation of what a scoring run is versus an Experiment run, could be tightened or moved after the actionable guidance.

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?

No output schema exists, and the description does explain key return fields (status, window, resolved_sessions) plus the object-store preview, which is genuinely useful. But the boilerplate about '@obj_123' object IDs sits uneasily with a result set that also has a preview, and the definition never explains the pipeline_experiment_id field it newly introduces as where the Metric Group is named, leaving the returned record shape partly ambiguous.

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 largely does: it explains that metric_group_id and session_id are mutually exclusive alternatives, and that limit is capped at 200 with no paging so an exactly-full result may be truncated. Only pipeline_name is handled tautologically ('Name of the pipeline'), which is the sole shortfall.

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 opening sentence gives a specific verb and resource: 'Lists scoring runs: a Metric Group's on one pipeline, or every one that judged a given Session.' It distinguishes two retrieval modes and routes single-run grid reading to get_experiment_run. The slight confusion between the tool name (list_experiment_runs) and the terminology 'scoring runs'/'pipeline_experiment_id' keeps it from a 5.

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?

The description states a clear selection rule between its own two modes ('Give metric_group_id for that group's runs, or session_id for ... that judged that Session') and points to get_experiment_run for grids. However, among a sibling set containing list_evaluation_runs, list_evaluation_tries, get_experiment_results and get_evaluation_run, only one alternative is named, so an agent gets no help choosing between the various list/get evaluation tools.

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