tradeoffs
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted conversation history and user-supplied arguments to generate output and save it to the file system.
- Ingestion points: The skill reads the current conversation history and takes a focus hint or path via
$ARGUMENTSinSKILL.md. - Boundary markers: No explicit delimiters or warnings to ignore instructions embedded in the conversation data are present.
- Capability inventory: The skill possesses file system capabilities, including directory creation (
mkdir -p) and file writing as described inSKILL.md. - Sanitization: The skill lacks explicit sanitization for user-provided paths or conversation content, but it implements a functional safeguard to avoid overwriting existing files by appending a numeric suffix to filenames.
- [SAFE]: The identified file writing and data processing behaviors are directly aligned with the skill's primary stated purpose of generating decision documentation within a project repository.
Audit Metadata