Conciseness

The body is dense and example-driven with no explanations of concepts Claude already knows; the opening paragraph restates the frontmatter description and the two input parameters are shown both as a table and as a JSON schema, minor duplication that could be trimmed. Otherwise every line carries tool-specific information.

Actionability

Fully executable guidance: a copy-paste-ready CLI invocation with JSON input, stdin/heredoc and --input-file variants, troubleshooting for the CLI not being found, and roughly twenty concrete RQL example queries covering set, call, create, destroy, reparent, scenes, prefabs, and editor state.

Workflow Clarity

For a batch/destructive skill, validation is explicitly present ("dry_run=true returns the receipt ... without changing anything", "All-or-nothing: every target and value is checked before anything is written", readonly-member error behavior), so the missing-validation cap does not apply. However, there is no explicit preview-then-apply sequence or failure-recovery loop (e.g. 'run dry_run first, review the receipt, then execute'), keeping it below a 5.

Progressive Disclosure

No bundle files exist, and the single-file body is well organized into How to Call, Troubleshooting, Input, and Output sections, with one-level-deep, clearly signaled cross-references to the rql-query skill and /unity-initial-setup. The body exceeds 50 lines and inlines a large example catalog, so it does not qualify for the simple-skill exception to score 5.

Specificity

The description names the domain ("live Unity scene") and several concrete capabilities ("bulk edits in one all-or-nothing, undoable call, saved to disk; dry_run previews"), going well beyond 1-2 actions. It falls short of comprehensive coverage since create/destroy/prefab operations are absent and the qualifiers all attach to a single core 'change' action.

Completeness

The 'what' is clear and concrete (atomic, undoable, persisted scene edits with preview), but there is no 'Use when...' clause or equivalent explicit trigger guidance, which caps completeness at 3 per the judging guidelines.

Trigger Term Quality

It includes some natural keywords a user would say ("Unity", "scene", "edits") but leans on tool jargon ("RQL set/call", "dry_run") and misses common variations like "Unity Editor", "GameObject", "modify", or "prefab".

Distinctiveness Conflict Risk

"Change the live Unity scene with RQL set/call" carves out a clear niche distinct from read-only querying, though there is minor overlap risk with a sibling rql-query skill in the same tool family.