Content
80%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 lean, command-driven workflow that is fully executable through the validation stage and appropriately compact for a skill whose job is delegation. Its main gap is the missing verification loop after the batch fixing is delegated — nothing tells Claude to confirm that the fixes actually succeeded before finishing.
Suggestions
Add post-fix verification steps: after the codacy-handler agent finishes, re-fetch Codacy findings to confirm they are resolved and run relevant tests before reporting completion.
Include a concrete example of the delegation prompt with real values (e.g., mode: 'backlog', categories: 'quality') and specify how the handler agent is spawned.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | At roughly 40 lines with a compact argument table, exact commands, and no background explanation of concepts Claude already knows — 'Check each of the following. On failure, print a clear error and stop' instructs without padding — the body matches anchor 5's 'lean and efficient; every token earns its place'. Not 4 because there are no instances of over-explanation that could be trimmed. | 5 / 5 |
Actionability | Every validation step gives an exact, copy-paste-ready command ('gh pr view --json number', 'op --version', 'git branch --show-current', 'git rev-parse --is-inside-work-tree'), matching anchor 4's 'concrete code or commands with minor gaps'. Not 5 because the delegation step is a placeholder template ('Run the codacy-handler workflow with: mode: `<resolved mode>`') with no concrete example prompt, and 'print a clear error' specifies no error content; not 3 because nothing is pseudocode and the commands are fully executable. | 4 / 5 |
Workflow Clarity | The sequence (parse arguments → validate preconditions → delegate) is clearly ordered with an explicit stop-on-failure checkpoint ('On failure, print a clear error and stop — do not spawn the agent'), but the skill's purpose is autonomously fixing findings — a batch code-change operation — and there is no verification after the agent runs (no re-check of findings, no tests, no diff review), so per the batch-operation cap it cannot exceed anchor 3 ('sequence present but checkpoints missing'). Not 4 precisely because the missing outcome-verification is more than a minor gap for batch fixes. | 3 / 5 |
Progressive Disclosure | The skill is under 50 lines, has no bundle files and no external references to organize, and its three well-labeled sections ('## Parse arguments', '## Validate preconditions', '## Delegate to agent') organize the content cleanly, so the simple-skill exception applies and anchor 5 is warranted. Not 4 because there is no content that should have been split out and no organization gaps. | 5 / 5 |
Total | 17 / 20 Passed |