Content
62%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 lean, well-organized interpretive reference that adds genuine non-obvious evidentiary judgment (e.g., 'A key embedded in a bundle may be public by design'). Its weaknesses are the absence of any sequenced workflow with checkpoints and the reliance on unnamed external skills for all executable detail.
Suggestions
Name the delegated skills and sources concretely — e.g. 'Reuse the android-static-analysis and ios-static-analysis skills' and identify where 'the live schema' is consulted — so the pointers are actionable rather than hand-wavy.
Add a minimal ordered procedure with explicit checkpoints, e.g. 1. fingerprint the app (hashes/IDs/signatures) 2. enumerate candidates (exported components, bridges, secrets) 3. confirm reachability before treating as finding 4. attribute captured traffic to app and session.
For each evidentiary limit, state what evidence would satisfy it — e.g. what 'authorized backend evidence and a baseline showing what the key grants' concretely looks like — to convert criteria into verifiable steps.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is ~25 dense lines with zero padding: it assumes competence ('Tool availability and argument names come from the live schema. Authorization comes from the RoE.') and never re-explains concepts Claude already knows; every token earns its place. | 5 / 5 |
Actionability | It names concrete tools (apktool, jadx, class-dump, Frida, TLS-pinning bypasses) and concrete artifacts to examine (exported components, intent filters, WebView bridges, trust store, root/jailbreak state), but gives no commands, examples, or steps — implementation is delegated to unnamed 'existing technique skills', so guidance describes rather than instructs. | 3 / 5 |
Workflow Clarity | The skill explicitly disclaims defining an order ('they do not define an engagement stage order') and none is given: the static→dynamic sections provide only a topical grouping with no sequenced steps or validation checkpoints, and criteria like 'claimed access needs authorized backend evidence and a baseline' are not framed as verify-steps. | 2 / 5 |
Progressive Disclosure | No references/, scripts/, or assets/ files exist, and this short (<50 line) self-contained note is cleanly organized into two clear sections — but its cross-references ('Reuse existing Android/iOS skills', 'the relevant existing technique skills', 'the live schema') are unnamed and unpathed, a minor navigation gap. | 4 / 5 |
Total | 14 / 20 Passed |