- Repository:
https://github.com/BeatAPI/beatapi-skill - Skill directory:
skills/beatapi-video - Upload ZIP:
dist/beatapi-video-skill.zipafternpm run submission:build - Skill name:
beatapi-video - Invocation:
$beatapi-video - License: MIT
- Product documentation:
https://docs.beatapi.io/ - API key creation:
https://beatapi.io/dashboard/apikeys
- Run
npm run contract:syncagainst the reviewed BeatAPI public contract. - Run
npm run verify. - Run the official host validator when the target directory provides one.
- Confirm the repository is public and the default branch is
main. - Confirm the Skill folder has no symlinks, generated caches, private media, credentials, or repository-only relative references.
- Test a fresh clone with every prompt in
evals/evals.json. - Confirm the listing distinguishes the two adapters: bundled BeatAPI MCP in
the complete Codex plugin, or the globally installed
beatapiCLI for a standalone Skill submission. - Capture the repository URL, one-sentence description, categories, and support/security links required by the directory form.
Name: BeatAPI Agent Toolkit
Short description: Use BeatAPI models, Social Data, and workflows.
Long description: Discover current text, image, and video models, inspect
Social Data actions, and run production workflows through BeatAPI. The Skill
uses the provider-neutral Search → Inspect → Run loop, preserves direct API and
asynchronous task semantics, and keeps API keys outside prompts. It prefers
BeatAPI MCP tools supplied by the host and otherwise uses the official
beatapi CLI.
Standalone prerequisite: A Skills-only submission does not include an MCP
server. Users therefore need Node.js and the globally installed beatapi CLI
unless their host already supplies compatible BeatAPI MCP tools.
Categories: Developer Tools, Data, Automation, AI Media
Attach or link:
- successful CI run;
- contract SHA from
contract/contract.lock.json; - official Skill validator output;
- clean secret scan;
- screenshots only if they contain no credentials or private media.