heuristic-eval

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill operates entirely within the local development environment, focusing on usability analysis of project assets. It does not request network access or execute shell commands.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface that could be targeted for indirect prompt injection.
  • Ingestion points: The skill processes user-provided arguments and a wide range of local project files including source code, screenshots, and documentation.
  • Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the instructions.
  • Capability inventory: Skill capabilities are limited to reading local files and writing a markdown report to docs/hci/heuristic-evaluation.md. It has no access to network or command execution tools.
  • Sanitization: No content sanitization is performed on the ingested data, as the agent is expected to analyze it as semantic context.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 04:52 PM
Security Audit — agent-trust-hub — heuristic-eval