The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs users to install the Render CLI by piping a remote shell script directly to the command line. Evidence: curl -fsSL https://raw.githubusercontent.com/render-oss/cli/main/bin/install.sh | sh in SKILL.md. This resource originates from the author's official GitHub repository and is a standard installation method for the platform's tooling.
[COMMAND_EXECUTION]: The instructions involve executing local shell commands to verify system state and prerequisites. Evidence: Commands such as git remote -v, render --version, and render whoami -o json are used to ensure the environment is correctly configured for deployment.
[INDIRECT_PROMPT_INJECTION]: The skill performs codebase analysis to generate deployment configurations, which constitutes a data ingestion surface. 1. Ingestion points: The agent reads repository files like package.json, requirements.txt, and go.mod (referenced in SKILL.md and references/codebase-analysis.md). 2. Boundary markers: No explicit delimiters or instructions to ignore embedded prompts are provided during the analysis phase. 3. Capability inventory: The skill can write configuration files (render.yaml), invoke platform-specific MCP tools (e.g., create_web_service), and execute shell commands for status verification. 4. Sanitization: The instructions do not specify sanitization or filtering of codebase content prior to its use in generating the deployment Blueprint.

render-deploy