Content
78%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 highly actionable, well-sequenced iOS red-team runbook with strong command-level guidance and a useful decision tree. Its main weakness is progressive disclosure: it is a monolithic inline document with no bundle files, so detail that would benefit from being split into reference files is all in SKILL.md.
Suggestions
Extract the Stage 3 secret-grep pattern catalog and Stage 4 ATS red-flag list into a references/ file (e.g. references/ios-secret-patterns.md) and link to it from the body, so the overview stays lean.
Trim the extended BoringSSL explanation in Stage 4 to the essential hook targets and preferred tools (SSL Kill Switch 2 / objection) to improve token efficiency.
Add an explicit validation/checkpoint note after Stage 2 unpacking and Stage 3 grep (e.g. 'confirm strings dump is non-empty before proceeding') to strengthen workflow feedback loops.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is command-driven and largely lean, but a few asides could be trimmed — e.g. the extended BoringSSL paragraph in Stage 4 and the repeated tool-install lines — without losing actionable value; not 5 due to those minor over-explanation instances, not 3 because the bulk is efficient practical guidance rather than concepts Claude already knows. | 4 / 5 |
Actionability | Nearly every stage is copy-paste-ready bash (plutil, codesign, class-dump, frida, objection, grep patterns) with concrete flags and inline objection commands, fully covering the common iOS red-team cases. | 5 / 5 |
Workflow Clarity | A clear Stage 0→6 sequence is paired with a decision-tree routing table and an anti-patterns guardrail list; the destructive/batch validation cap does not apply to this recon pipeline, but per-stage validate-then-proceed feedback loops are implicit rather than explicit, so it sits at 4 rather than 5. | 4 / 5 |
Progressive Disclosure | The skill is well-organized into staged sections but, at ~287 lines with no references/scripts/assets bundle, all detail (ATS red-flag catalog, cert-pinning bypass variants, secret-grep patterns) is inlined rather than split into one-level-deep reference files; not 4 because there are no external file references to signal navigation, not 2 because section structure is genuinely clear and cross-skill chains are well signaled. | 3 / 5 |
Total | 16 / 20 Passed |