The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of API calls and shell commands through the one CLI and its workflow engine, which supports bash steps and Node.js code modules.
[EXTERNAL_DOWNLOADS]: The one mem command initiates a one-time download of Postgres binaries (~52MB) from the vendor to support local database and vector search functionality.
[DATA_EXFILTRATION]: The skill is designed to access and process data from 250+ external platforms. It can read sensitive information and has the capability to write this data to local files or transmit it to other APIs.
[PROMPT_INJECTION]: As an orchestration tool ingesting data from external APIs, the skill identifies the risk of indirect prompt injection. Ingestion points: Data enters through actions execute responses and webhook payloads. Boundary markers: The CLI uses structured JSON output (--agent). Capability inventory: Workflow steps support bash commands, code modules, and file-write operations. Sanitization: Built-in mitigations include Handlebars pipes (| shell, | json, | url) and POSIX shell-quoting helpers (q) to ensure untrusted data is safely handled within commands.

one