Debugging
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- COMMAND_EXECUTION (LOW): The script
scripts/find-polluter.shexecutesnpm test(or user-defined alternatives) on local files. While this is the intended functionality of a debugging tool, it involves executing code found within the repository. - PROMPT_INJECTION (LOW): The skill is subject to indirect prompt injection (Category 8) as it requires the agent to ingest and analyze untrusted data from error logs and test outputs.
- Ingestion points:
scripts/find-polluter.sh(stdout/stderr from tests) andreferences/systematic-debugging.md(guidance to read all error messages). - Boundary markers: Absent; there are no specific instructions or delimiters used to prevent the agent from obeying malicious instructions embedded in debug logs or test results.
- Capability inventory: The agent has the ability to execute local shell commands via the provided script and read files throughout the project.
- Sanitization: No sanitization or filtering is performed on the data captured from test outputs before the agent processes it.
Audit Metadata