The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references an official installer located at https://cdn.buildin.ai/buildin-cli/install. This domain is the distribution endpoint for the vendor's CLI tool.\n- [REMOTE_CODE_EXECUTION]: The skill provides instructions for downloading and executing the buildin CLI installer. It mitigates the inherent risks of remote execution by requiring the agent to obtain explicit user approval before proceeding and recommending integrity checks via SHA-256 hashes or signatures.\n- [COMMAND_EXECUTION]: The skill uses the buildin command-line utility to perform API operations. It enforces security best practices by instructing the agent to use environment variables for authentication instead of command-line flags, preventing sensitive tokens from appearing in process logs or shell history.\n- [INDIRECT_PROMPT_INJECTION]: As the skill retrieves content from external sources (Buildin workspace pages and databases), it possesses an attack surface for indirect prompt injection.\n
Ingestion points: Data is ingested via commands such as buildin page get, buildin markdown get, and buildin search defined in SKILL.md.\n
Boundary markers: No specific boundary markers or "ignore embedded instructions" delimiters are implemented for the fetched content.\n
Capability inventory: The skill is capable of executing subprocess commands (buildin CLI) and writing to the local file system.\n
Sanitization: The skill relies on retrieving content as Markdown for inspection but does not provide automated sanitization or filtering of the remote data before it enters the agent context.

buildin-cli