replit-deploy-integration

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides legitimate instructions for deploying applications to Replit, using standard configuration formats and following best practices for managing production secrets via Replit's native secrets feature.\n- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch and process data from remote endpoints via curl for health checks, which establishes an attack surface for untrusted data ingestion.\n
  • Ingestion points: Health check endpoints at user-defined deployment URLs (e.g., $DEPLOY_URL/health).\n
  • Boundary markers: No specific delimiters or instructions to ignore embedded content within the remote response are provided in the prompts.\n
  • Capability inventory: The skill utilizes Bash(curl:*), Read, Write, and Edit tools to manage project files and verify deployments.\n
  • Sanitization: The skill uses jq for field extraction from the JSON response, providing structural validation but no content sanitization for potential downstream prompt context.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:48 AM
Security Audit — agent-trust-hub — replit-deploy-integration