process-interviewer
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to automatically scan existing workspace context (including project folders, existing skills, and
CLAUDE.mdfiles) to answer its own questions and form recommendations. This data ingestion surface allows external, untrusted content within the project to potentially influence the agent's reasoning or the output of the generated skills and plans. - Ingestion points: The
SKILL.mdfile (Interview Rule 2) explicitly directs the agent to scan existing skills, reference files,CLAUDE.mdfiles, and project folders to retrieve context. - Boundary markers: No specific delimiters or "ignore instructions" markers are defined for processing the ingested workspace data.
- Capability inventory: The skill is capable of writing new
SKILL.mdfiles, reference documents, and plan files to the local workspace (Phase 5). - Sanitization: No sanitization or validation logic is specified for the data retrieved from the workspace before it is interpolated into the agent's context or the final output artifacts.
Audit Metadata