Content
40%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.
The body is a well-organized catalogue stub, but as executable guidance it is weak: the install instruction has no concrete command, the only code block just opens a URL, and there is no validation that the upstream skill was actually installed and is invocable. Duplicating the frontmatter description and source metadata also wastes tokens.
Suggestions
Replace the `open https://...` code block with a concrete, executable step for obtaining the upstream skill (e.g., a `git clone` or copy command into the agent's skills directory), or explicitly state that no local workflow exists and the entry is catalogue-only.
Remove the duplication of the frontmatter description in 'What it does' and the repeated upstream URL/category in 'Source'; one clean statement plus a single clearly-labeled link suffices.
Add a verification checkpoint, such as confirming the installed skill appears in the agent's skill list or triggers on 'pptx'/'powerpoint', so users can detect a failed install before relying on it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short but padded with duplication: the frontmatter description is repeated verbatim in "What it does", and "Source" restates the upstream URL and category already in the frontmatter. This matches anchor 3 ('mostly efficient but includes some unnecessary explanation or could be tightened') — no concepts Claude doesn't know are explained, so it is not a 2. | 3 / 5 |
Actionability | The only instruction is to "install the upstream bundle into your active agent's skills directory" with no command for doing so, and the sole code block (`open https://github.com/...`) merely opens a URL while punting with "Inspect the upstream README for exact paths". This is high-level hints without executable steps — anchor 2 — and lacks even the partial concrete code of anchor 3. | 2 / 5 |
Workflow Clarity | A rough two-step sequence exists (install the upstream bundle, then ask the agent to invoke the skill), but the critical install step is undefined, there is no verification that the skill was installed or loaded, and no error-recovery guidance. Anchor 2 ('rough sequence present but many gaps; validation absent') fits; it is not a 1 because the sequence is coherent. | 2 / 5 |
Progressive Disclosure | No bundle files exist (no references/, scripts/, or assets/), and the body is well-sectioned (What it does / Source / How to use) with a single one-level pointer to the upstream repository. This sits between anchors 3 and 4: structure is good, but the essential pointer is buried inside a code block and source information is duplicated from the frontmatter rather than cleanly signaled. | 4 / 5 |
Total | 11 / 20 Passed |