debugging-wizard
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data through log analysis and error message parsing.
- Ingestion points: SKILL.md defines a workflow for gathering error messages and stack traces from the execution environment.
- Boundary markers: Absent. No delimiters are specified to isolate untrusted log data from the agent's instructions.
- Capability inventory: The skill identifies and uses various system tools including language debuggers (pdb, node, dlv) and version control (git).
- Sanitization: No filtering or sanitization protocols are defined for the input data.
- [METADATA_POISONING]: A discrepancy exists between the author name in the skill's metadata (Jeffallan) and the provided author context (hosseinmirzapur).
- [REMOTE_CODE_EXECUTION]: The skill documentation references the installation of external libraries and the use of package managers. Evidence: Mentions of 'npm install' for troubleshooting and references the 'go-spew' library on GitHub.
Audit Metadata