kernel-audit-trace
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a protocol for architectural tracing and system analysis. It utilizes standard read-only tools (
Read,Grep,Glob) to inspect local source code. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because it reads and processes external files from a user's workspace as directed in
SKILL.md. However, it lacks capabilities to write to the file system or communicate over a network, and it explicitly instructs the agent not to implement code, which effectively mitigates the risk of malicious instructions within the audited source code influencing the environment.
Audit Metadata