claymation-ad

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from external sources, which creates a potential surface for indirect prompt injection. However, this is heavily mitigated by the skill's design and referenced security protocols.
  • Ingestion points: The skill fetches content from user-provided URLs (product pages), reads .agents/product-context.md, and processes user-supplied ad copy and product images.
  • Boundary markers: The skill explicitly references references/agent-security.md and mandates a check to verify that all fetched/pasted input was treated as data rather than instructions. It requires the agent to quote and report any embedded instructions rather than obeying them.
  • Capability inventory: The skill's primary capabilities are prompt generation and writing research findings to .agents/product-context.md. It does not have access to high-privilege shell commands, arbitrary network POST operations, or sensitive system directories.
  • Sanitization: The skill follows references/agent-security.md, which requires HTML-escaping for all interpolated values in emitted markup or code and validates URLs to https: or relative paths.
  • [PROMPT_INJECTION]: The skill contains standard security instructions to protect against prompt injection, specifically requiring the agent to disregard instructions found within untrusted data and treat it only as data to be analyzed.
  • [CREDENTIALS_SAFE]: The skill includes a dedicated security reference (references/agent-security.md) that forbids persisting secrets, credentials, or tokens in files or output. It provides clear instructions for the agent to flag potential credentials in input for rotation rather than reproducing them.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 04:06 PM