agentic-eval

Pass

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill describes patterns that ingest untrusted data into iterative refinement loops, which represents an indirect prompt injection surface.
  • Ingestion points: The task parameter in Pattern 1/2 and the spec parameter in Pattern 3 (SKILL.md) ingest external requirements.
  • Boundary markers: No delimiters or instructions to ignore embedded commands are present in the provided templates.
  • Capability inventory: The CodeReflector pattern references a run_tests function which implies the execution of agent-generated code.
  • Sanitization: The patterns do not include logic for sanitizing or validating generated code before execution.
  • [SAFE]: The skill consists entirely of markdown documentation and Python code snippets illustrating standard AI engineering best practices. No actual executable scripts, hardcoded credentials, network exfiltration patterns, or persistence mechanisms were detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 24, 2026, 03:48 PM
Security Audit — agent-trust-hub — agentic-eval