architect
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill operates by analyzing project-specific data such as source code, manifests, and existing documentation, which inherently presents a surface for instructions embedded in those files to be processed by the agent.
- Ingestion points: As specified in
references/_shared/orchestration/reality-check.md, the agent is directed to inspect the current branch, HEAD, working tree, code paths, manifests, and migrations. - Boundary markers: The instructions do not provide explicit delimiters or instructions to treat data from the project files as non-executable content.
- Capability inventory: The
SKILL.mdworkflow includes the creation and modification of project artifacts such as Architecture Decision Records (ADRs) and plans. - Sanitization: There are no instructions for sanitizing or validating the content extracted from external project files before it is used in the decision-making process.
Audit Metadata