Content
100%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.
An exemplary skill body: a complete, tightly sequenced workflow with an explicit feedback loop, anti-gaming guard ("never reword it to get a PASS"), a fixed output contract, and an escalation stop — all in about 26 lines with no wasted tokens. Nothing material is missing for the skill's scope.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~26-line body contains zero padding and no explanations of concepts Claude already knows; even the opening scoping sentence earns its place by delimiting the boundary with the sibling fast-review skill. Every token instructs. | 5 / 5 |
Actionability | For an instruction-only skill the guidance is copy-paste ready: the exact question ("Is this change safe to merge?"), the materials that answer it, a four-section reviewer output template, the counting rule, the union-fix rule with tally marking (2/3), and the escalation rule. Not 4: there are no gaps in the executable instructions. | 5 / 5 |
Workflow Clarity | Five explicitly sequenced steps with validation checkpoints throughout: the vote is the gate, step 4 requires every MUST-FIX "fixed, or answered with evidence", and step 5 is an explicit feedback loop (BLOCK → change work → re-run the same question → stop after a second BLOCK). The output-format block doubles as a reviewer checklist. Not 4: checkpoints and the error-recovery loop are explicit, not implicit. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines, single-purpose, and needs no external references — the rubric's simple-skill exception applies, and the numbered steps plus the code-block template are well organized with no content that belongs in a separate file. | 5 / 5 |
Total | 20 / 20 Passed |