Content
57%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 works well as a README-style overview: concise, cleanly sectioned, with concrete example invocations. But as a skill it is a launcher for an external repo — the operational content (command catalog, scoring dimensions, state management) that should be bundled in reference files is absent, leaving execution details under-specified.
Suggestions
Bundle a command reference file (e.g. references/commands.md) cataloging all 23 commands with their arguments, instead of showing only 4 examples and deferring the rest to the external repo.
Define the five scoring dimensions used by transcript "analyze" and the answer-scoring workflow, either inline or in a referenced file, so the coaching logic is actually executable.
Document the coaching_state.md lifecycle — where it lives, when it is read/updated, and how kickoff output flows into later prep/mock/negotiation commands — to give the multi-command workflow explicit sequencing.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean with no explanations of concepts Claude already knows; commands are shown once with brief outcomes. Minor trimming possible: the Overview restates the frontmatter description, "23 total commands" repeats it again, and the Examples section partially overlaps "When to Use This Skill" and "What It Covers" — so efficient but not every token earns its place. | 4 / 5 |
Actionability | Concrete invocations exist ("npx skills add dbhat93/job-search-os", "/coach prep Stripe Senior PM", "/coach analyze"), but operational detail is missing: only 4 of the "23 commands" are shown, the "five dimensions" scoring is never defined, and the actual instructions live in the external GitHub repo rather than anywhere in this skill. Some concrete guidance, but key details needed to execute the skill's purpose are absent — not a 2 because the given commands are executable and each example states its outcome. | 3 / 5 |
Workflow Clarity | A rough sequence is implied (install → "/coach kickoff" → per-task commands), and state persistence via "coaching_state.md" is mentioned. But the ordering and relationship between the 23 commands is never laid out, and there are no checkpoints (e.g., how prep feeds mock drills, when to update the storybank). Sequence present but gaps and implicit checkpoints — not a 2 because the install and kickoff entry path is explicit and coherent. | 3 / 5 |
Progressive Disclosure | Sections are well-organized and the SKILL.md stays a genuine overview, with no nested references. However, the bundle contains no reference files at all — the command catalog, scoring rubric, and state-file format that clearly belong in separate one-level-deep files are simply missing (deferred to an external repo), so structure exists but the detail layer is absent rather than well-signaled. | 3 / 5 |
Total | 13 / 20 Passed |