detect-bullshit-tests
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads test files and production code in full to perform audits. Because these files are untrusted data, they could contain malicious instructions designed to manipulate the agent. The skill does not implement boundary markers or sanitization to protect against such injections.
- [DYNAMIC_EXECUTION]: The skill executes project-specific formatters and tests. This requires running local scripts or binaries which could be malicious if the repository has been compromised.
- [COMMAND_EXECUTION]: The skill performs shell operations including
git diffandgit commit. These powerful capabilities, especially when used in the autonomous--automode, could be exploited to make unauthorized changes to the codebase if the agent's instructions are subverted.
Audit Metadata