external-repo-guidelines

Pass

Audited by Gen Agent Trust Hub on Jun 19, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill creates a surface for Indirect Prompt Injection by explicitly instructing the agent to prioritize instructions found in external, untrusted files (such as CONTRIBUTING.md, AGENTS.md, or .github/CONTRIBUTING.md) over its own default guidelines and patterns.
  • Ingestion points: The agent is directed to read from the ## Contributing Guidelines section in CONTEXT.md, which is populated with data from external, potentially attacker-controlled repositories.
  • Boundary markers: There are no instructions to use delimiters or ignore embedded malicious directives within the external content.
  • Capability inventory: The skill encourages the agent to follow repo-specific conventions for coding, formatting, and executing test commands (e.g., ./gradlew test), which could be leveraged to execute arbitrary malicious scripts.
  • Sanitization: The skill lacks any sanitization logic or warnings to the agent regarding the validation of external instructions before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 19, 2026, 05:58 PM
Security Audit — agent-trust-hub — external-repo-guidelines