Content
76%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-organized reference: all commands are executable with complete flag documentation, plus good staleness and recovery guidance. The main gaps are the absence of validation guidance around the destructive `clean` command and minor content that could be offloaded to reference files.
Suggestions
Add a validation checkpoint before destructive/batch clean operations, e.g. "Run `node .gitnexus/run.cjs status` first and confirm the repo/index shown is the intended target; never run `clean --all` without listing repos via `list`".
Move the wiki flag table and the npm 11.x `node.target is null` workaround into a references file (e.g. TROUBLESHOOTING.md) to keep SKILL.md a lean overview.
Include a verify-after-clean step (e.g. re-run `status` to confirm the repo is unregistered) to close the feedback loop on the destructive workflow.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and command-focused with almost no concept re-explanation, but a few tokens could be trimmed: the hook-design rationale ("to avoid blocking the agent for up to 120s and risking KuzuDB corruption") and the inline npm 11.x version-specific workaround, which the guidelines penalize outside a dedicated old-patterns section. | 4 / 5 |
Actionability | Every command is fully executable and copy-paste ready ("node .gitnexus/run.cjs analyze", "npx gitnexus analyze"), with complete flag tables, defaults ("default: minimax/minimax-m2.5", "default: 3"), and a concrete recovery command for the missing-runner case. | 5 / 5 |
Workflow Clarity | Sequencing is good — explicit "When to run" triggers, an "After Indexing" checklist, and a verification step ("Read gitnexus://repo/{name}/context to verify the index loaded") — but the destructive/batch operations lack validation: `clean` (deletes the index) and `clean --all` (batch) have no pre-flight check or confirm-what-will-be-deleted guidance, which caps this dimension at 3 per the destructive-operations rule. | 3 / 5 |
Progressive Disclosure | Sections are well organized (Commands, After Indexing, Troubleshooting) and self-contained with no nested references, but at ~85 lines — beyond the 50-line simple-skill exception — peripheral material like the wiki flag table and the npm 11.x issue workaround could live in separate reference files. | 4 / 5 |
Total | 16 / 20 Passed |