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

Annotations declare write-but-not-destructive, non-idempotent behavior; the description adds substantive traits beyond that — it reveals that omitted finishing touches are removed (a replace-in-full semantic), that prior clips are preserved except as requested, and that it does not trigger a planner or render. These are exactly the behavioral details an agent can't infer from the hints.

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 and most sentences carry a real rule, but the block is very dense and the trailing 'Arguments: {...}' JSON example largely restates structure the agent already receives as a full schema, which is mild redundancy. Efficient overall, but not tightly structured.

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 large nested plan payload with no output schema and rich per-field $defs descriptions, the tool-level description covers the non-obvious essentials: unit conventions, replace-on-omit behavior, clip preservation, and the approval workflow. The remaining field-level richness (panels, recipe, bonuses, music) is handled by the schema descriptions, so little is genuinely missing.

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?

Top-level schema descriptions are absent (0% coverage), but the description compensates with high-value semantics that the structured fields do not convey: clip start/end are SOURCE seconds while finishing touches (texts, overlays, subtitles, zooms) use FINAL montage seconds, along with the speed range and the constraint that only selected job_ids are valid. The source-vs-final-seconds distinction is critical and is added here, not by the schema.

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 and resource — 'Save an external assistant's structured edit plan' — and scopes it with 'without invoking a planner or rendering', which distinguishes it from generation-oriented siblings like kelpo_generate or kelpo_preview_editing. It is clearly a persistence operation on an existing plan, though it does not name a sibling explicitly.

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 gives concrete workflow context: send the complete plan, show the full recap, and request explicit approval before render_editing. That establishes the precondition and ordering. What it lacks is an explicit alternative-based rule (e.g. 'use this instead of kelpo_preview_editing when...'), so it is strong context but not full when/when-not routing.

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