Content
92%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 tight, operational skill body: fully executable command syntax, an explicitly validated workflow with feedback loops and blocker handling, and well-organized self-contained sections. The only improvement opportunity is trimming the mild repetition between the workflow steps, Decision Rules, and Red Lines.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense, imperative, and teaches nothing Claude already knows, but there is minor duplication: the retry/resubmit/ack definitions appear in both workflow step 5 and Decision Rules, and the no-restart rule appears in both the intro and Red Lines. Not 5 because not every token strictly earns its place; not 3 because there is no genuine padding or over-explanation. | 4 / 5 |
Actionability | Exact executable command syntax with parameter placeholders throughout: `ccb trace <id>`, `ccb queue --detail <agent|all>`, `ccb pend --inbox --detail <agent>`, `ccb repair retry <job_id|attempt_id>`, `ccb cancel <job_id>`. Every repair path is copy-paste ready. | 5 / 5 |
Workflow Clarity | The 8-step Core Workflow sequences trace-diagnosis, classification, in-flight cancellation, gated repair execution, and post-repair re-verification with explicit stop-and-report blocker conditions — a complete validate/fix/re-check feedback loop around destructive lineage mutations. | 5 / 5 |
Progressive Disclosure | No bundle files exist and none are needed; the ~85-line body contains only essential decision logic split cleanly across Core Workflow, Decision Rules, Artifact Rules, and Red Lines with no nested or buried references. Navigation is trivial. | 5 / 5 |
Total | 19 / 20 Passed |