vale
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious activities or security vulnerabilities were identified. The skill follows established safety guidelines for agent operations.
- [PROMPT_INJECTION]: The presence of injection patterns like "ignore previous instructions" in "references/evals.yaml" is part of a security evaluation framework. These patterns are used to confirm the agent correctly identifies and blocks malicious attempts during testing.
- [DATA_EXFILTRATION]: The instructions in "SKILL.md" and the configuration in "references/contract.yaml" explicitly require the redaction of credentials and private data, preventing accidental exposure.
- [COMMAND_EXECUTION]: The skill involves the use of Vale for linting workflows but includes safeguards against unauthorized destructive commands and restricts the scope of actions to confirmed goals.
Audit Metadata