Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-organized, actionable, and token-efficient offensive-BLE playbook that scores highly on conciseness, actionability, and structure. The main gap is the lack of explicit validation checkpoints around the destructive OTA/firmware-write operations.
Suggestions
Add an explicit validation checkpoint before OTA/DFU writes: confirm operator approval + recovery plan, then verify the firmware image checksum and post-write device integrity before declaring success.
Define the referenced dump_gatt.py (used in the Evidence section) as an actual script in scripts/ or inline, so the evidence command runs as written.
Add a short retry/rollback loop for the Just Works pairing step (re-pair with a rotated BD_ADDR on failure) to match the OPSEC guidance with a concrete feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and task-focused across all phases with executable snippets, a UUID attack table, and OPSEC notes, and it avoids explaining concepts Claude already knows like what BLE is or how libraries work. | 3 / 3 |
Actionability | It provides copy-paste-ready bleak and gatttool code, concrete bluetoothctl/sriffle/ubertooth commands, and a specific UUID-to-attack mapping table, giving direct executable guidance rather than vague direction. | 3 / 3 |
Workflow Clarity | The five-phase sequence and Evidence hashing give a clear progression, but destructive operations like OTA/DFU firmware writes lack explicit validate-before-write / verify-after feedback loops, which the rubric caps at 2. | 2 / 3 |
Progressive Disclosure | No bundle files exist; the skill is a single well-organized SKILL.md with clearly labeled phase sections and one-level-deep external reference URLs (Sniffle, bleak) that are not nested, satisfying the simple-skill exception. | 3 / 3 |
Total | 11 / 12 Passed |