test-feedback
Pass
Audited by Gen Agent Trust Hub on Mar 29, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it instructs a background agent to process and act upon potentially untrusted data from test failure logs.
- Ingestion points: Reads test output content from user-specified file paths or the default /tmp/output (as defined in SKILL.md).
- Boundary markers: The instructions lack explicit delimiters or instructions to the background agent to ignore or escape any natural language instructions that might be embedded within the test output logs.
- Capability inventory: The spawned background agent is granted extensive capabilities, including reading source code, writing fixes to files, and executing shell commands like make test and make test-integration (SKILL.md).
- Sanitization: There is no evidence of sanitization or validation of the input file content before it is passed to the background agent for analysis and execution.
- [COMMAND_EXECUTION]: The background agent is directed to execute shell commands (make test and make test-integration) based on its analysis of the parsed test failures.
- [DATA_EXFILTRATION]: The skill accesses the local file system to read files provided by the user or from common temporary locations like /tmp/output.
Audit Metadata