doca-pcc-ztr-rttcc-algo
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill assists users in modifying C source code and executing build/debug commands (using
meson,ninja, anddpacc). This creates an attack surface where malicious instructions could be embedded in external data processed by the agent. - Ingestion points: User-provided C source files (specifically
rp_rtt_template_dev_main.c), build scripts, and environment logs/outputs used during the debugging workflow. - Boundary markers: The skill does not provide explicit delimiters or warnings to help the agent distinguish between untrusted user data and its own instructional logic.
- Capability inventory: The agent is guided to perform file modifications and support shell command execution for building and testing network software.
- Sanitization: There are no instructions for validating, escaping, or filtering the external content or logs processed during the task workflows.
Audit Metadata