nda-triage

Pass

Audited by Gen Agent Trust Hub on Oct 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides Python scripts (scripts/nda_screener.py and scripts/nda_checklist.py) intended to be executed by the agent. These scripts use the Python standard library to perform text analysis and do not request elevated privileges or network access.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external text files (NDAs). This creates a vulnerability surface where malicious instructions could be embedded in the text being triaged.
  • Ingestion points: The nda_screener.py and nda_checklist.py scripts ingest data from external files provided via command-line arguments.
  • Boundary markers: There are no explicit delimiters used in the scripts to isolate the input text from the agent's instructions, although the scripts themselves use static regex rather than LLM-based analysis.
  • Capability inventory: The skill has the capability to read local files and write results to the local file system. It does not possess network or shell-injection capabilities.
  • Sanitization: No specific sanitization or filtering for prompt injection patterns is implemented for the ingested text.
  • [SAFE]: The skill performs its primary function (legal document triage) using local, verifiable scripts. No obfuscation, persistence mechanisms, or credential harvesting patterns were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 29, 2026, 11:37 AM