Content
63%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 dense, highly actionable reference whose per-workflow tool sequences, parameters, and pitfall annotations are genuinely valuable. Its weaknesses are redundancy (Known Pitfalls and Quick Reference duplicate earlier sections), a filler closing section, no error-recovery loops, and a monolithic single-file layout where split reference files would serve progressive disclosure.
Suggestions
Remove the redundant 'Known Pitfalls' section and the vacuous 'When to Use' section; fold anything not already covered into the relevant workflow's Pitfalls list.
Split per-domain detail (e.g., issues/PRs/repo-admin workflows with their parameter and pitfall tables) into one-level-deep reference files like references/issues.md and references/pull-requests.md, keeping SKILL.md as an overview with the Setup and safety rules.
Add one concrete example invocation with real argument values per workflow (e.g., a GITHUB_CREATE_A_PULL_REQUEST call with owner/repo/head/base filled in) to close the actionability gap, and brief fix-retry guidance for the named failure modes (422/404/409).
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly lean, high-value Rube-specific detail (tool slugs, params, error codes) that Claude cannot know, but it includes padded redundancy: the 'Known Pitfalls' section repeats the per-workflow pitfalls nearly verbatim, the Quick Reference table restates the tool sequences, and the closing 'When to Use' section ('This skill is applicable to execute the workflow or actions described in the overview.') is pure filler. It is above 2 because the padding is duplication rather than explanation of concepts Claude already knows, but below 4 because three sections could be trimmed or removed outright. | 3 / 5 |
Actionability | Gives exact tool names with Required/Optional sequencing, concrete parameters (owner, repo, head, base, merge_method), and specific failure modes (422 on invalid base/head, 404 on unprotected branches, per_page max 100, 384KB search index limit) — genuinely executable guidance. Not 5 because there are no example invocations with actual argument values (e.g., a sample RUBE_MANAGE_CONNECTIONS or GITHUB_CREATE_A_PULL_REQUEST call), leaving minor gaps for the most common cases. | 4 / 5 |
Workflow Clarity | Each workflow has a 'When to use' gate and a numbered, labeled tool sequence, and destructive/batch operations do include validation checkpoints (verify mergeable status immediately before merge, require explicit user confirmation for merge/delete, check CI runs before merging, iterate pagination until empty) — so the destructive-op cap at 3 does not apply. Not 5 because there are no error-recovery feedback loops (e.g., what to do after a 422 or a failed auth check) beyond naming the failure modes. | 4 / 5 |
Progressive Disclosure | No bundle files exist (no references/, scripts/, or assets/), and the entire ~230-line reference lives inline in SKILL.md. Section organization is good, but per-domain API details and the quick-reference table are exactly the content that belongs in one-level-deep reference files, placing it at the 'some structure but could be better organized; content that should be separate is inline' anchor. Not 4 because there is no split at all, not 2 because navigation within the file is clear and well-sectioned. | 3 / 5 |
Total | 14 / 20 Passed |