product-marketing

Pass

Audited by Gen Agent Trust Hub on Jul 8, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill features an 'auto-draft' workflow that ingests content from local codebase files such as README.md, package.json, and existing marketing documentation. This constitutes a surface for indirect prompt injection if these files contain instructions intended to manipulate the agent's behavior. However, the risk is addressed by the workflow requiring a manual user review of the generated draft before it is saved to the filesystem.
  • Ingestion points: Local codebase files (README, landing pages, marketing copy, package.json, etc.).
  • Boundary markers: No explicit delimiter-based isolation is defined for the ingested data.
  • Capability inventory: The skill has file read and write access within the project directory, specifically targeting the .agents/ and .claude/ folders.
  • Sanitization: No automated sanitization or filtering of ingested file content is performed; the skill relies on the user to review and correct the output.
  • [COMMAND_EXECUTION]: The skill performs only standard file read and write operations within the project directory. It does not invoke shell commands, execute subprocesses, or perform any network operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 8, 2026, 05:45 AM
Security Audit — agent-trust-hub — product-marketing