test-harness
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to ingest and analyze untrusted data from the local project repository (such as source code, test files, and documentation) to verify behaviors and perform testing tasks.
- Ingestion points: The agent is instructed to read various project files to understand application logic, verify bug fixes, and update test coverage maps (SKILL.md, references/harnesses.md).
- Boundary markers: The instructions do not specify the use of delimiters or 'ignore embedded instructions' warnings when the agent processes content from the repository.
- Capability inventory: The skill involves powerful capabilities, including executing shell commands via subprocesses to run test suites (Vitest, Playwright, Supabase CLI) and performing file writes to modify code and documentation (SKILL.md, references/harnesses.md, references/rules-from-bugs.md).
- Sanitization: There are no explicit instructions for sanitizing or validating external file content before the agent uses it to generate commands or update repository files.
Audit Metadata