fw-core
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a dispatcher for user-submitted firmware tasks. * Ingestion points: Natural language user prompts describing firmware objectives (referenced in SKILL.md and task-typing.md). * Boundary markers: Includes strict instructions to ask for clarification and avoid guessing. * Capability inventory: The skill itself performs environment checks but delegates implementation to specialized sub-skills; it requires confirmation for destructive actions (e.g., eMMC flashing). * Sanitization: Employs an explicit refusal and hand-off policy for unsupported technologies like Rust no_std firmware.
- [COMMAND_EXECUTION]: The skill uses basic shell commands (uname, make, idf.py) to verify development environment compatibility. These are standard, low-risk diagnostic operations typical for firmware build tools.
Audit Metadata