The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @dokploy/cli package from the NPM registry to enable deployment capabilities. This is a standard dependency for the skill's primary function.
[COMMAND_EXECUTION]: The bootstrap_dokploy.sh and inspect_project.mjs scripts execute shell commands to manage the CLI installation, verify Node.js versions, and query local Git metadata (status, branches, remotes).
[DATA_EXFILTRATION]: The skill is designed to send project metadata, build configurations, and environment variables to the Dokploy console at http://211.47.74.86:3000. This is the intended endpoint for deployment operations managed by the vendor.
[CREDENTIALS_UNSAFE]: The skill requests and utilizes a DOKPLOY_API_KEY for authentication. It provides instructions for secure handling via environment variables and includes explicit warnings against printing sensitive tokens in logs or chat responses.
[SAFE]: Indirect Prompt Injection surface analysis:
Ingestion points: Reads local project files including package.json and Dockerfile via scripts/inspect_project.mjs.
Boundary markers: The extracted metadata is structured into JSON for agent consumption rather than being directly interpolated into executable instruction blocks.
Capability inventory: The skill possesses file-read (metadata) and network-write (deployment) capabilities.
Sanitization: Data is processed using JSON parsing and regex matching for specific deployment parameters (e.g., ports, framework names).

onestack