improve-website

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions advise the user to install additional sub-skills via npx skills add wondelai/skills/<slug> --global if a specific methodology phase is missing. These resources are hosted under the author's repository.
  • [COMMAND_EXECUTION]: The skill suggests the use of the npx command-line tool to manage extensions, which involves executing code outside the immediate skill environment.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, such as live website copy, CSS, and visitor input, which creates a surface for indirect prompt injection.
  • Ingestion points: File SKILL.md under the 'Intake' section, item 7 ("URLs, live copy, CSS, analytics screenshots").
  • Boundary markers: Absent. The skill does not define specific delimiters to isolate external content from the agent's instructions.
  • Capability inventory: The skill writes to local files in the docs/ directory and invokes other functional skills (cro-methodology, ux-heuristics, etc.) to process data.
  • Sanitization: No validation or sanitization routines are specified for the user-provided assets.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 12:04 PM
Security Audit — agent-trust-hub — improve-website