project-development-mindset
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external repository data, which can serve as a vector for indirect prompt injection attacks.
- Ingestion points: Instructions in SKILL.md direct the agent to orientation phases that involve inspecting "source code, configuration, schemas, routes, tests, logs, and generated artifacts."
- Boundary markers: The skill includes instructions to "distinguish authority from evidence" and to avoid "copying blindly," which provides a conceptual boundary, though no structural data delimiters are defined for processing untrusted strings.
- Capability inventory: The skill authorizes the agent to implement code changes, execute project scripts and tests, and update documentation, creating a capability set that could be exploited if malicious instructions are embedded in the files being read.
- Sanitization: While the skill mandates keeping secrets out of durable context, it lacks specific technical sanitization or escaping mechanisms for the repository data it interpolates into its operational context.
Audit Metadata