architect-worker
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes an attack surface for indirect prompt injection by design, as both roles are instructed to follow documentation stored in the workflow/ directory.
- Ingestion points: The Architect role reads workflow/BACKLOG.md (which may contain external product requirements), while the Worker role reads SPEC.md, PLAN.md, TASKS.md, and REVIEW.md from the workflow// directory.
- Boundary markers: The skill does not explicitly define delimiters or "ignore instructions" warnings for the content read from these files.
- Capability inventory: The Worker role has the capability to write production application code and execute shell commands for verification (test, lint, build). The Architect role similarly executes verification commands and governs the workflow files.
- Sanitization: No explicit sanitization of the input data is mentioned. The role-based separation and mandatory code review process act as procedural mitigations against malicious instructions that might be embedded in specifications.
Audit Metadata