fix-failing-tests
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from a user-provided test output file, which could theoretically contain malicious instructions.
- Ingestion points: Step 1 reads content from a file path provided by the user or a default path at
/tmp/output. - Boundary markers: There are no explicit delimiters or specific instructions for the agent to ignore instructions embedded within the test output file.
- Capability inventory: The skill can execute shell commands (
rg,make test) and modify files to apply fixes. - Sanitization: No sanitization is performed on the test output content before it is passed to a subagent for diagnosis.
- [COMMAND_EXECUTION]: The skill uses command-line tools to identify test failures and verify that code changes solve the reported issues.
- Evidence: Step 2 uses
rg(ripgrep) to search for specific error markers like 'FAIL', 'panic', and 'DATA RACE' within the output file. - Evidence: Step 4 executes
make testto validate applied fixes, with a limit of 3 retry attempts if tests continue to fail.
Audit Metadata