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 catalogue stub rather than a functional skill: it is structurally tidy and concise in length, but its guidance is mostly a pointer — the only executable step is opening the upstream GitHub page, and the critical install instruction is left undefined. Duplication of the description and provenance wastes tokens, and there is no usable local content for PDF tasks.
Suggestions
Replace the vague install instruction with a concrete, executable command (or step sequence) for installing the upstream bundle into the agent's skills directory, so the workflow's pivotal step is actionable.
Add or inline a minimal quick-start (e.g., a one-line example of extracting text or filling a form) so the skill provides some usable guidance locally instead of only deferring to the external repository.
Remove redundancy by dropping either the provenance blockquote or the Source section (they repeat the same URL) and cutting the "What it does" section that restates the frontmatter description verbatim.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short (~25 lines) but contains unnecessary repetition: the "What it does" section restates the frontmatter description verbatim, and provenance is given twice (the blockquote "Curated from Anthropic's official skills repository" and the Source section with the same upstream URL). This fits anchor 3's "mostly efficient but includes some unnecessary explanation" rather than 4. | 3 / 5 |
Actionability | Only one executable command is provided ("open https://github.com/...") while the core instruction — "install the upstream bundle into your active agent's skills directory" — has no steps, path, or command, and "Inspect the upstream README for exact paths" defers everything else. This matches anchor 2's minimal concrete guidance missing the specific steps to execute; the single real command keeps it above 1. | 2 / 5 |
Workflow Clarity | A rough sequence exists (open URL → install upstream bundle → invoke skill), but the pivotal install step is undefined with no concrete commands and no validation checkpoints, fitting anchor 2's "rough sequence present but many gaps; steps poorly defined". The sequence is coherent enough to stay above anchor 1; the undefined key step keeps it below anchor 3's merely-missing-checkpoints level. The skill is non-destructive, so the validation cap does not apply. | 2 / 5 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent), and the body itself has clean, well-organized sections with a single clearly signaled one-level pointer (the upstream URL) — matching anchor 4's good structure with clearly signaled references. It does not reach 5 because the entire substantive content is deferred to an external repo rather than organized bundle files, leaving the local SKILL.md a pointer with no self-contained overview. | 4 / 5 |
Total | 11 / 20 Passed |