Ai (destructive)

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

Annotations already declare destructiveHint=true and readOnlyHint=false, but the description adds real behavioral value: the mandatory plan/confirm handshake and the fact that clear_transcription is irreversible (asymmetry vs clear_audio_classification). It does not cover permission requirements or what the snapshot/backup actually preserves.

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 purpose, then a tightly formatted action list, then the workflow rule. The closing line 'Calls known in advance (a sequence, the same change on several items): ONE batch_run' is terse to the point of being cryptic but does not waste space.

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, nested-schema tool with an output schema this is close to complete: purpose, available actions, per-action params, confirm flow, and bulk alternative are all present. The remaining gap is auth/permission expectations and the reversibility of the audio-classification clear, which is left unstated.

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% for the top-level params, so the baseline is 3; the description does add per-action parameter names (clip, folder, nested) that the loose 'op' object with additionalProperties=true does not enumerate, which is a genuine gain. However, the stated '* = required' convention is never applied — no asterisk appears on any listed param — leaving requiredness per action ambiguous.

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 verb 'Remove' plus the enumerated action list ('clear_audio_classification', 'clear_transcription') makes the resource scope concrete, going beyond the vague 'ai objects' of the first sentence. It distinguishes itself from sibling 'batch_run' for bulk sequences, though it does not contrast with 'ai_read'/'ai_write' directly.

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?

It states the two-phase calling protocol explicitly (first call returns plan + confirm_token, repeat with the token to apply) and routes pre-planned multi-item work to a sibling: 'ONE batch_run.' What's missing is any guidance on when to choose ai_delete over ai_write for these same objects.

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