Content
88%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.
This is a strong, highly operational audit procedure: exact search patterns, executable fixes, a verify-before-report loop, compound-severity cross-referencing, and a concrete output contract. The main weaknesses are minor redundancy in the tool-use mandate and invocation examples, and references to companion skill files that are not clearly signaled or bundled.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and operational — exact grep patterns, severity tables, tunable thresholds, and anti-drift corrections like 'There is no synchronous per-change plist rewrite to look for'. Minor trimming opportunities remain: the tool-use mandate is restated three ways, the five Invocation Examples are near-identical, and the Scope section repeats the frontmatter description — matching 'Efficient; minor instances of over-explanation' rather than the lean level-5 anchor. | 4 / 5 |
Actionability | Every pattern ships exact grep patterns, executable fix code (e.g., 'try data.write(to: url, options: [.atomic, .completeFileProtection])' and the setResourceValues shadowing snippet), a graphviz location decision tree, a complete output-format template, health-scoring criteria, and output limits. This matches 'Fully executable; copy-paste ready code or commands; specific examples cover the common cases'. | 5 / 5 |
Workflow Clarity | Phases 1-5 are clearly sequenced, each detection pattern follows a Search-Verify-Fix loop with explicit verification ('use Read to verify the surrounding context before reporting'), the Storage Map must be presented before proceeding, and False Positives plus Output Limits handle edge cases. This matches the level-5 anchor with explicit validation checkpoints and error-recovery guidance; as a read-only audit it is not subject to the destructive-operation cap. | 5 / 5 |
Progressive Disclosure | A well-organized single-file skill (no references/scripts/assets bundle exists) with clear section headers and one-level-deep pointers to related skills ('axiom-data (skills/storage.md)', 'axiom-security (skills/file-protection-ref.md)'). It falls short of the level-5 anchor because those pointers are not clearly signaled as links and some reference-shaped material (the decision tree, cross-auditor overlap notes) is inlined in a ~320-line file rather than split out. | 4 / 5 |
Total | 18 / 20 Passed |