Content
77%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-structured operational skill with strong error-recovery and verification guidance for side-effect operations. Its weaknesses are redundancy (duplicated action table and repeated query-projection boilerplate) and progressive disclosure: the full action contract is inlined rather than split out, and the one reference to deeper documentation is a dangling path.
Suggestions
Drop the action summary table (or reduce it to action names only) since the per-action sections already carry the full contract, and state the `query` projection rules once in the 'Query Content Projection' section instead of repeating them inside every action.
Move the complete action contract into a reference file (e.g. references/actions.md) and keep only the fast-path usage, call shape, and safety rules in SKILL.md, signaling the reference clearly.
Fix or remove the dangling 'See docs/media-server-query.md' reference — the file is absent from the skill bundle, so either bundle it or describe the projection compatibility limits inline.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient with no beginner-concept padding, but the action summary table ('| `activity.backdrops` | 读取服务近期媒体的背景图...arguments: `limit`, `remote` |') duplicates the per-action sections that follow, and the `query` projection sentence ('统一内容投影，如 {"fields":["identity","path"]}；省略或 fields=null 保留原内容...') is repeated verbatim ~10 times instead of being defined once and referenced. | 3 / 5 |
Actionability | Fully executable guidance throughout: exact bash invocations ('python skills/mediaserver-operation/scripts/mp-mediaserver.py call --server "living-room" --action activity.latest --arguments \'{"limit":20}\''), complete per-action argument types, defaults, required markers (e.g. '`item_id*` (string)'), enums, and a copy-paste-ready external MCP JSON example. | 5 / 5 |
Workflow Clarity | Clear sequenced workflow with explicit validation and feedback loops: the fast path ('If multiple servers remain ambiguous, the failed call lists every valid server name. Reuse that list for the next direct call'), batch error correction ('correct every reported field together instead of retrying one field at a time'), and pre-confirmation plus post-verification for side-effect actions ('Before `library.scan`, confirm the exact server and scan scope', 'After a scan or refresh, verify with `activity.latest`, `items.detail`, or the smallest relevant library query'). | 5 / 5 |
Progressive Disclosure | Section headers and the script path reference are well organized, but ~130 lines of the complete action contract ('This is the complete Media Server Operation action contract... a field name ending in `*` is required') are inlined in SKILL.md where a separate reference file would fit better, and the single pointer to deeper detail ('See `docs/media-server-query.md` for the mappings and compatibility limits') targets a file that does not exist in the bundle. | 3 / 5 |
Total | 16 / 20 Passed |