The Agent Skills Directory

REMOTE_CODE_EXECUTION (HIGH): Detected piped shell execution for the Render CLI installer. Automated scan found curl -fsSL https://raw.githubusercontent.com/render-oss/cli/main/bin/install.sh | sh. While this is the official installer, the render-oss organization is not on the trusted list, and the execution method is a known high-risk vector. The severity is adjusted to HIGH as this pattern is associated with the primary purpose of the skill.
PROMPT_INJECTION (LOW): High surface for indirect prompt injection via project manifests. Ingestion points: package.json, requirements.txt, pyproject.toml, and Dockerfile (referenced in references/codebase-analysis.md). Boundary markers: Absent; the agent is not instructed to treat project file content as untrusted. Capability inventory: The skill can create web services, cron jobs, and execute shell commands. Sanitization: Absent; extracted build and start commands are used without validation.
COMMAND_EXECUTION (LOW): The skill utilizes local shell commands for service validation and log viewing. references/deployment-details.md lists commands like render blueprints validate and render services.
EXTERNAL_DOWNLOADS (LOW): References downloading the CLI and project-specific dependencies. Mentions curl for the CLI and various package managers (npm, pip, go) for dependencies during the build process.

render-deploy