agentsclimarketplace
Sign in

Tokenlab api integration

Skill hedging8563/tokenlab-skills/skills/tokenlab-api-integration

Integrates TokenLab chat, image, audio, video, and other API families into code with runnable examples, model discovery, public contract checks, and agent-first recovery paths. Use when the user wants to add TokenLab APIs to an app or script.From its SKILL.md

Install
npx -y skills add hedging8563/tokenlab-skills --skill tokenlab-api-integration

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

3 things to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
  • runs commandsInstructs the agent to run 3 commands, including `GET /v1/models` and 2 more.
  • fetches URLsInstructs the agent to fetch 1 URL, including https://api.tokenlab.sh/llms.txt.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

3.7 KB, 713 tokens by cl100k_base, as published. Nobody here has run it

TokenLab API Integration

Built for runnable integration work for TokenLab chat, image, audio, video, and other API families across coding agents.

What this skill should deliver

  • A minimal runnable example using the fewest moving parts possible.
  • The exact base URL, auth shape, install command, and environment variables required to run the example.
  • A concise note on when to stay on the OpenAI-compatible path versus switching to a native Anthropic or Gemini route.
  • For non-chat APIs, a model discovery or contract check before hardcoding request shape details.
  • A concrete default model choice that is plausible on TokenLab, not a generic placeholder.
  • A short explanation of the agent-first recovery path when the model, endpoint, or route guess is wrong.

Preferred approach

  1. Clarify the user's goal, inputs, and required deliverable.
  2. Read references/usage-notes.md before acting.
  3. Produce one concrete output before adding explanation.
  4. Use the following operating rules:
  • Start with the smallest working example before introducing abstractions or helper layers.
  • State the base URL explicitly and keep the environment setup copy-pasteable.
  • When model selection is open, show how to discover models through /v1/models or https://api.tokenlab.sh/llms.txt instead of hardcoding one option.
  • For non-chat model selection, prefer GET /v1/models?recommended_for=<scene> where <scene> is one of image, video, music, 3d, tts, stt, embedding, rerank, or translation.
  • Before retrying a failed non-chat request, read GET /v1/models/:model and align with the public contract, including supported_operations, supported_parameters, request_endpoint, request_shape_mode, and recommended_request.
  • Use native Anthropic or Gemini examples only when the request explicitly needs provider-specific behavior.

Output format

  • One short intro sentence explaining what the example does.
  • One runnable code block only.
  • One shell setup block showing both dependency install and the exact environment variable export.
  • One short model discovery note.
  • One short routing note explaining when to stay on the OpenAI-compatible path and when a response header or provider-specific feature suggests a native Anthropic or Gemini route.

Avoid

  • Do not return pseudo-code when runnable code is expected.
  • Do not hide required environment variables, auth headers, or base URLs.
  • Do not over-claim pricing, speed, or compatibility without grounding it in a concrete example or source.
  • Do not claim an exact platform-wide model count; say "hundreds of models" unless the current API response is being quoted directly.
  • Do not silently drop unsupported non-chat fields. If removing a field would change user intent, safety, billing, or response guarantees, surface the contract error and fail closed.

Inputs

  • Natural-language user request
  • Referenced files or URLs
  • Existing project context, if available

Outputs

  • A concrete deliverable, recommendation, or implementation result
  • Short notes on assumptions, caveats, or next actions when needed

Edge Cases

  • If required inputs are missing, state exactly what is missing.
  • If the request only partially matches this skill, handle the matching portion and clearly scope the rest.
  • If a risk, safety, or compliance concern appears, surface it before producing the final output.

What ships with it: 3 files

3.5 KB alongside SKILL.md

agents/

references/

Gives 0 of the 12 instructions most video audio skills give in 713 tokens

Counted across 619 of the 725 authors here whose files we hold, read 2026-09-06

  • Read product marketing context firstin 13 of 619, across 7 files
  • Define the core visual thesis in one sentencein 11 of 619, across 3 files
  • Break the concept into 3 to 6 scenesin 11 of 619, across 3 files
  • Render the smallest working version firstin 11 of 619, across 3 files
  • Start with a low-quality smoke test renderin 11 of 619, across 3 files
  • Add captions for accessibility and engagementin 11 of 619, across 5 files
  • Write the scene outline before writing codein 11 of 619, across 3 files
  • Specify subject, action, camera, style, and moodin 11 of 619, across 5 files
  • Decide what each scene provesin 10 of 619, across 2 files
  • Export one clean thumbnail framein 10 of 619, across 2 files
  • Pick the right tool for the jobin 10 of 619, across 4 files
  • Run the test suite before proposing a fixin 8 of 619, across 7 files

Said here and by no other author read

  • Clarify the user goal inputs and required deliverable
  • Read references/usage-notes.md before acting
  • Produce one concrete output before adding explanation
  • Start with the smallest working example
  • State the base URL explicitly
  • Discover models through endpoint or llms.txt instead of hardcoding

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.