marketing-context

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions do not contain any prompt injection or safety bypass patterns. It remains focused on its marketing purpose.\n- [SAFE]: The included validation script scripts/context_validator.py uses only Python standard libraries and performs local file processing without network access or unsafe dynamic execution.\n- [SAFE]: No hardcoded credentials, sensitive file access, or persistence mechanisms were identified in any of the skill files.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies a functional requirement to process project files which may contain external content.\n
  • Ingestion points: Codebase repository files such as README.md, landing pages, and package.json are read to draft the context.\n
  • Boundary markers: The skill does not define specific delimiters to separate ingested file content from the agent's instructions.\n
  • Capability inventory: The skill is capable of reading project files and writing the results to marketing-context.md.\n
  • Sanitization: There is no explicit sanitization or filtering of the ingested project data.\n- [SAFE]: The skill does not perform any external downloads or remote code execution, relying entirely on provided local files and project repository contents.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 05:04 AM
Security Audit — agent-trust-hub — marketing-context