create-website

Pass

Audited by Gen Agent Trust Hub on Sep 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to recommend a shell command to the user for installing missing components. Specifically, it suggests the command npx skills add wondelai/skills/<slug> --global (in SKILL.md) when a sub-skill is not available. This resource originates from the skill author's repository.
  • [INDIRECT_PROMPT_INJECTION]: The skill maintains project state by reading local markdown files from the docs/ directory to resume workflows across different sessions.
  • Ingestion points: The agent reads docs/CREATE-WEBSITE-PLAN.md and other journey artifacts to summarize the state and determine the next phase (in SKILL.md).
  • Boundary markers: There are no instructions for the agent to use delimiters or specific safety prompts to ignore potential instructions embedded in the ingested documentation files.
  • Capability inventory: The skill can perform file writes, invoke other agent skills, and suggest the execution of shell commands to the user.
  • Sanitization: No explicit sanitization, filtering, or validation process is defined for the content retrieved from the docs/ folder.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 6, 2026, 11:35 AM
Security Audit — agent-trust-hub — create-website