Skills
skills/blogic-cz/blogic-marketplace/andocs/Gen Agent Trust Hub

andocs

Fail

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill provides instructions to install the Bun runtime via a remote script from its official site using a shell pipe. It also uses bunx to execute the andocs package directly from a remote registry.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided documentation, including diagrams and HTML snippets, which could contain instructions for the agent.
  • Ingestion points: User documentation directories (e.g., docs/, content/) and Markdown files.
  • Boundary markers: Fenced code blocks with language tags like mermaid and html-preview are used to delimit external content.
  • Capability inventory: The skill can execute local shell commands through the Bun runtime and start a local web server as documented in SKILL.md.
  • Sanitization: The skill documentation states that interactive previews and prototypes are rendered within sandboxed iframes to mitigate risks.
Recommendations
  • HIGH: Downloads and executes remote code from: https://bun.sh/install - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 9, 2026, 01:35 PM