debrief
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from incident descriptions and codebase artifacts, which represents a potential surface for indirect instructions.
- Ingestion points: User-provided incident context ($ARGUMENTS), git history logs, PR metadata, and CI/CD failure logs retrieved via the GitHub CLI tool.
- Boundary markers: The prompt instructions attempt to wrap provided context in <incident_context> tags to delineate it from system instructions.
- Capability inventory: The skill possesses capabilities to write to the local file system (docs/debriefs/), execute shell commands (git and gh), and interact with the user via specific developer tools.
- Sanitization: There is no evidence of explicit sanitization or filtering of the content retrieved from external logs or user input before it is interpolated into the generated documents.
Audit Metadata