dev-loop
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface for indirect prompt injection by instructing the agent to ingest and act upon untrusted project data.
- Ingestion points: The skill requires the agent to inventory and inspect repository structures, build definitions, CI workflow files, and recent failure logs (SKILL.md, audit-playbook.md).
- Boundary markers: The instructions lack boundary markers or explicit warnings to the agent that audited project files should be treated as untrusted or potentially adversarial content.
- Capability inventory: The skill empowers the agent to implement code optimizations, modify CI configurations, and install guardrails, providing a mechanism for injected instructions to reach high-privilege operations (SKILL.md).
- Sanitization: There are no requirements for sanitizing, validating, or filtering the content retrieved from the audited files before the agent processes or implements changes based on that content.
Audit Metadata