chief-loop
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files that could contain malicious instructions designed to influence the behavior of the builder or decision-support agents.
- Ingestion points: Data is read from files within the
_goal/,_contract/, and.chief/_rules/directories, as well as the_plan/_todo.mdfile (SKILL.md). - Boundary markers: The instructions do not specify the use of clear delimiters or boundary markers (e.g., XML tags or triple quotes) when interpolating file contents into prompts for sub-agents.
- Capability inventory: The skill performs file read and write operations (Markdown files in the
.chief/directory) and makes tool calls to spawn other agents. - Sanitization: The skill does not mention any validation, filtering, or escaping of the content read from these files before it is used to direct agent activity.
Audit Metadata