bruno-test-writer

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The SKILL.md file contains defensive instructions designed to prevent indirect prompt injection. It explicitly tells the agent to treat ingested content as untrusted data and to ignore any commands (like 'ignore previous rules') embedded within that data.
  • [DATA_EXFILTRATION]: The skill includes comprehensive rules against the exposure of credentials and PII. It instructs the agent to use placeholders or secret management helpers instead of hardcoding values from sample responses.
  • [COMMAND_EXECUTION]: The file scripts/test_generate_tests.py utilizes subprocess.run to execute the local generate_tests.py script. This usage is restricted to the test suite for verifying code generation logic and does not involve executing untrusted or remote commands.
  • [INDIRECT_PROMPT_INJECTION]: While the skill processes untrusted input (API samples and documentation), it mitigates the associated risks by including clear boundary markers and safety constraints that prioritize local configuration over external data.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 02:16 PM
Security Audit — agent-trust-hub — bruno-test-writer