Content
73%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-structured, actionable instruction skill with an excellent bounded-retry workflow and clear validation checkpoints. Its main weakness is repetition of the same constraints (no git, iteration cap) across multiple sections, which inflates token use without adding clarity.
Suggestions
State the no-git/no-PRs rule and the max_signal_iterations cap once in the Invocation Contract, and reference rather than restate them in Workflow, Signal Processing, and Guidelines to remove ~30% of redundant tokens.
Consider moving the Signal Processing section's markdown template and the TODO Structure example into a reference file (or collapsing them) to tighten the core flow.
Consolidate the DON'T list so each constraint appears exactly once, since several (no git, no scope creep) currently overlap with prose in earlier sections.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient and assumes Claude's competence with git/worktrees, but the no-git/no-PRs rule and the signal-iteration cap are each restated 3-4 times across sections, which is padding that could be tightened. | 3 / 5 |
Actionability | Concrete invocation contract (named inputs/outputs), explicit FAILURE reasons, and specific DON'Ts give mostly executable guidance; templated signal invocation and absence of literal commands are the minor gaps. | 4 / 5 |
Workflow Clarity | A clear 6-step sequence with an explicit validation gate (signal validation before unit tests) and a bounded feedback loop with explicit FAILURE conditions and iteration cap — matches the top anchor. | 5 / 5 |
Progressive Disclosure | Well-organized into clearly headed sections with no nested references and no bundle files; at ~90 lines it slightly exceeds the under-50-lines simple-skill threshold, keeping it just below a clean 5. | 4 / 5 |
Total | 16 / 20 Passed |