Skills
Navigate
skills/sisodiabhumca/agent-skills/knowledge-base-gap-mapper/Gen Agent Trust Hub

knowledge-base-gap-mapper

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The script reads untrusted JSON data from the file system, which could contain malicious instructions designed to influence the agent's behavior.
  • Ingestion points: run.py reads input data from a user-specified path using input_path.read_text().
  • Boundary markers: The skill does not implement delimiters or safety instructions to separate data from commands.
  • Capability inventory: The skill has file system read and write permissions via the run.py script.
  • Sanitization: No validation or filtering is performed on the ingested data.
  • [DATA_EXFILTRATION]: The skill implementation permits reading and writing to arbitrary file paths on the host system.
  • Evidence: The --input and --output arguments in run.py allow the agent to read any file and write results to any location accessible by the process.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 11:39 AM