get_sanitized_dataset

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

Goes well beyond the annotations: it discloses the output-directory sandbox (POISONDASH_OUTPUT_DIR, Downloads default), that relative paths resolve there, that path escape attempts including symlinks are refused, and the overwrite guard. These are exactly the non-obvious side effects an agent needs before invoking a destructiveHint=true tool.

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?

Three sentences, each carrying distinct load (what it produces, the precondition, the filesystem safety rules). Purpose is front-loaded and the sandbox constraint follows immediately.

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 destructive write tool with no output schema, this covers prerequisites, path safety, and overwrite behavior. Minor omission: no mention of the resulting file's naming, size, or how errors (e.g., failed precondition) are surfaced, though these are largely inferable.

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?

With 50% schema coverage, the description compensates well: it explains output_path resolution and its containment rule, and clarifies overwrite semantics ('won't overwrite unless overwrite=true'). It doesn't clarify what 'threshold' controls beyond the schema's own text or note experiment_id's role, leaving a small 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?

States a specific verb ('Download') and a precisely scoped resource ('the experiment's suspect dataset with flagged poisoned samples removed, as a .npz file'), which distinguishes it cleanly from the sibling get_dataset (the unfiltered dataset). Format and re-training intent are both named.

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 an explicit precondition ('Requires the experiment to be completed with per-sample scores present'), which tells the agent when this will fail. It does not name the alternative get_dataset or explain when the raw version is preferable, so it falls short of a 5.

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