publish-site

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill orchestrates site deployment using standard command-line utilities including git for versioning, gh for repository management, and provider-specific tools like wrangler and netlify. It also uses python3 for local serving and curl for live URL verification.
  • [EXTERNAL_DOWNLOADS]: Fetches and utilizes official deployment tools from well-known services, specifically Cloudflare's wrangler CLI and the netlify-cli package, to handle the transfer of build assets to hosting infrastructure.
  • [INDIRECT_PROMPT_INJECTION]: The skill implements a verification step that retrieves the first 30 lines of the live site content.
  • Ingestion points: The curl -sS command in SKILL.md fetches external content into the terminal output.
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are used when the content is retrieved.
  • Capability inventory: The skill possesses terminal execution capabilities, network access, and file system write access via git.
  • Sanitization: Potential risk is mitigated by truncating the output to the first 30 lines via head -30, limiting the volume of untrusted data processed.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 10:16 AM
Security Audit — agent-trust-hub — publish-site