Content
76%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.
The content is a lean, tool-rich reference cheat-sheet that is highly token-efficient and mostly actionable, but it lacks an explicit sequenced workflow with validation checkpoints and has minimal internal organization.
Suggestions
Add a brief numbered workflow with validation/verification checkpoints (e.g., confirm session token validity after AiTM, verify token scope after device-code flow) to lift workflow clarity above the destructive-skill cap of 3.
Provide one or two concrete command invocations (e.g., a sample evilginx3 phishlet launch or TokenTactics device-code request) to move actionability from playbook-level to copy-paste executable.
Break the monolithic code block into labeled sub-sections (Credential spraying / AiTM / Device code / Payloads / OAuth / Social engineering) to improve navigability within the single file.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is an extremely dense cheat-sheet where every token earns its place and assumes Claude's competence, with no over-explanation of what phishing or MFA is. | 5 / 5 |
Actionability | Names concrete tools (evilginx3, Modlishka, TokenTactics, AADInternals, gophish), payload types (lnk/iso/宏/HTA/OneNote), and bypass methods, giving mostly actionable guidance, but stops short of copy-paste command sequences. | 4 / 5 |
Workflow Clarity | Techniques are presented as a flat unordered list with only an implicit chaining hint at the end and no explicit validation checkpoints; for a destructive/attack skill, the missing validation feedback loop caps this at 3. | 3 / 5 |
Progressive Disclosure | A short (<50 line) self-contained single-section cheat sheet with no nested references, which is appropriate for a simple skill, though it is one undivided code block that could use internal sub-headers. | 4 / 5 |
Total | 16 / 20 Passed |