craft-content-modeling

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions for the agent to use CLI commands via the ddev craft utility. These commands are necessary for administrative tasks such as synchronizing project configuration, rebuilding indices, and performing content migrations.
  • Evidence: ddev craft project-config/rebuild, ddev craft entrify/categories, ddev craft up, and ddev craft project-config/touch are referenced throughout SKILL.md and references/infrastructure.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill suggests using the WebFetch tool to retrieve external documentation when local reference files are insufficient. This creates an attack surface where instructions embedded in external web content could potentially influence the agent's logic or behavior.
  • Ingestion points: Retrieval of external documentation via the WebFetch tool as suggested in SKILL.md.
  • Boundary markers: Absent. The skill does not provide instructions for the agent to differentiate between its core guidelines and any instructions found within fetched external content.
  • Capability inventory: The agent has the ability to execute terminal commands (ddev craft) and modify system files, providing a path for potential exploitation if malicious instructions are ingested.
  • Sanitization: Absent. There are no guidelines for validating or sanitizing the content fetched from external URLs.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 06:37 AM
Security Audit — agent-trust-hub — craft-content-modeling