chief-retro
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes the
git logcommand to extract commit history for the milestone or batch being reviewed. This is a standard operation for retrospective analysis. - [PROMPT_INJECTION]: The skill processes untrusted data which presents a surface for indirect prompt injection:
- Ingestion points: Reads project files (
_plan/_todo.md,_goal/*.md,_contract/*.md) and execution logs (_report/*.md), as well as git commit messages. - Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when interpolating this data into the prompt.
- Capability inventory: The agent has the capability to write retrospective reports to
.chief/_report/and create or update rule files in.chief/_rules/. - Sanitization: No programmatic sanitization of ingested content is performed; however, the skill implements a mandatory human-in-the-loop checkpoint, requiring the user to approve any proposed rule updates before they are written to the filesystem.
Audit Metadata