artifact-reuse-pattern
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs agents to maintain persistent repository documents (such as
implementation_plan.mdandwalkthrough.md) by updating them in-place with task results. - Ingestion points: Data is written to repository files based on agent activity and task outcomes across multiple turns.
- Boundary markers: The provided pattern and examples do not use delimiters or instructions to prevent the agent from potentially obeying malicious instructions embedded in the data being processed and stored.
- Capability inventory: The skill relies on file-writing and content-replacement tools to mutate the state of files within the repository.
- Sanitization: No guidance is provided for sanitizing or validating external or generated content before it is persisted to the canonical artifacts.
Audit Metadata