Content
83%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 exceptionally concise and actionable with strong command coverage, but as a destructive offensive operation it lacks explicit mid-workflow validation checkpoints and feedback loops, capping workflow clarity despite an otherwise excellent Step 1–6 structure.
Suggestions
Add explicit validation checkpoints between steps — e.g., after Step 2 confirm a client has associated to the rogue AP (check mana log / associated station) before issuing deauth or proceeding to portal capture.
Insert a feedback loop after portal/MITM capture: verify the credential was actually written (check mana_creds.txt / bettercap log) before recording evidence and tearing down.
Cross-link the deauth step more tightly to the deauth-pmf validation step (verify PMF state) as a precondition rather than a parenthetical note, to make the destructive-action gate explicit in-line.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and command-driven with no padding of concepts Claude already knows; every code block and comment (e.g., mana_loud semantics, portal-template list) earns its place. | 5 / 5 |
Actionability | Provides copy-paste-ready hostapd-mana, dnsmasq, wifiphisher, and bettercap invocations with real flags and justified placeholders, covering the common open-AP, PSK-clone, portal, and MITM cases. | 5 / 5 |
Workflow Clarity | Steps 1–6 are clearly sequenced and the top-level RoE gate is strong, but this destructive/offensive workflow lacks mid-process validation checkpoints and feedback loops (e.g., confirm client association before deauth, verify the portal was reached before declaring capture), which per the guidelines caps a destructive skill at 3. | 3 / 5 |
Progressive Disclosure | Single-file skill with clear section headers and a one-level-deep References section pointing to sibling skills and external tools; no nested references and no bundle files to over-split, though some content (portal templates, MAC-randomization defeat) is inlined where a reference could be considered. | 4 / 5 |
Total | 17 / 20 Passed |