manage-brand

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTIONNO_CODE
Full Analysis
  • [NO_CODE]: This skill consists of markdown instructions and YAML configuration files with no executable code or scripts included.
  • [EXTERNAL_DOWNLOADS]: The skill configures a Model Context Protocol (MCP) tool hosted at https://apikit.simplified.com/mcp. This connection is directed to the official API for Simplified, a well-known design and marketing platform, to facilitate the management of brand kits.
  • [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface because its core workflow involves analyzing external websites and documents to populate brand context documents.
  • Ingestion points: Instructions in SKILL.md direct the agent to analyze "websites, supplied documents, and approved messaging" for brand evidence.
  • Boundary markers: The skill does not employ explicit delimiters like XML tags or "ignore instructions" wrappers for the ingested content.
  • Capability inventory: The agent is authorized to create and update persistent brand records via the api_updateContextDocument, api_buildBrandKit, and api_createBrandKit tools.
  • Sanitization: The skill relies on logic-based guardrails (e.g., "Do not invent positioning", "Present a concise proposed delta") rather than programmatic sanitization to ensure data integrity.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 09:51 PM
Security Audit — agent-trust-hub — manage-brand