render-migrate-from-heroku

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to accept/paste Heroku config output and connection strings and to call update_environment_variables / use pg_credentials, which requires the LLM to include secret values (env var values, DB URLs, tokens) verbatim in tool calls or generated output.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.80). The majority of the links are official Render or well-known Git hosting/documentation pages, but the raw.githubusercontent.com link points to an install.sh that is recommended to be curl|sh — a direct remote shell script execution vector that is high-risk if run uninspected or if the source is compromised.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill includes runtime installation and runtime MCP endpoints that fetch and execute remote content: the install script fetched and executed via curl (https://raw.githubusercontent.com/render-oss/cli/main/bin/install.sh) and the MCP server endpoint used at runtime to provide app/config data (https://mcp.render.com/mcp), both of which can directly change agent behavior or execute code.