context-files

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill describes an architecture for ingesting external project files (e.g., AGENTS.md, .hermes.md) into the agent's context, which constitutes an indirect prompt injection surface.
  • Ingestion points: System files such as .hermes.md, AGENTS.md, CLAUDE.md, and SOUL.md as described in SKILL.md.
  • Boundary markers: The documentation highlights a multi-layer security pipeline including prompt injection scanning and character-length truncation to manage ingested data.
  • Capability inventory: The skill is instructional and does not provide active tools or scripts that execute system-level commands beyond the provided documentation examples.
  • Sanitization: Describes an automated scanning layer that validates content for UTF-8 compliance and malicious prompt patterns before loading.
  • [COMMAND_EXECUTION]: Includes standard shell command examples for user-side configuration and diagnostics.
  • Evidence: The skill provides examples of 'ls -la', 'cat', and 'ln -sf' for symbolic link management and troubleshooting file loading issues.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 07:24 AM
Security Audit — agent-trust-hub — context-files