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

Annotations declare write (readOnlyHint false), openWorld true, non-idempotent, non-destructive, and the description does not contradict that. Beyond annotations it adds substantial behavioral detail: providers searched in parallel, commercial-licence filtering, the real-file judging criteria (watermarks, motion, crop survival), grading/grain matching, and the licence line on the item. It does not cover retry/cost or how failures are reported, keeping it below a 5.

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?

The leading 'PROJECT-SCOPED:' and 'WRITE:' tags front-load the most important facts, which is good. But the core sentence is a single breathless run-on packed with parentheticals and arrows that is genuinely hard to parse. Structure is present at the top and then collapses.

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 complex, multi-provider, write-side tool with no output schema, the description covers what is searched, what survives filtering, what is produced, and the zero-library fallback. It also notes that the project identity is returned. Coverage is strong, with minor gaps around max/providers and failure reporting.

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 only 20%, so the description must carry weight. It explains the moments fields well (query = what the shot SHOWS, word/at, purpose, duration_s, kind, archival) and the place=false semantics. But it leaves max, providers, moments.id, and query_variants unexplained, so it only partially compensates for the coverage 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?

The description states a specific verb+resource: it sources stock B-roll and cuts the winner in as a broll fx item, and it is self-described as 'the B-roll director.' It also distinguishes itself from research-only behavior via 'place=false researches only.' However, it does not name its close siblings (research_broll, find_footage, add_stock_media) so an agent must infer the boundaries rather than being told them.

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: use it when you want footage sourced AND cut in, and place=false when you only want research. The 'No library live' fallback (returns a shoot list) is real conditional guidance. But there is no explicit when-not or routing to research_broll / find_footage / search_stock, so the agent is left to infer.

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