API

Quick Start

Upload a skill in one command — no account required. The response gives you a shareable URL and a one-liner install command.

1. Upload

curl -X POST https://skillb.in/api/skills \
  -H "Content-Type: text/markdown" \
  --data-binary @SKILL.md

2. Response includes

{
  "url": "https://skillb.in/s/aBc12345",
  "install": "curl -fsSL https://skillb.in/install/aBc12345 | sh"
}

3. Install

curl -fsSL https://skillb.in/install/aBc12345 | sh

Base URL

https://skillb.in

All endpoints below are relative to this base. Responses are JSON unless noted otherwise.

Authentication

Most read endpoints work without authentication. Creating skills linked to your account, managing tokens, and accessing private skills require a Bearer token.

Authorization: Bearer <your-token>

How to get a token:

• Sign in with GitHub on the web UI
• Your session API token is shown on the homepage when signed in
• For durable tokens: create one via POST /api/tokens (see below)

{
  "name": "my-ci-token",
  "expires_in_days": 90    // optional, null = never expires
}

{
  "token": "a1b2c3d4...",   // full token — only shown once
  "id": "uuid",
  "name": "my-ci-token",
  "prefix": "a1b2c3d4",
  "expires_at": "2026-08-10T00:00:00.000Z",
  "created_at": "2026-05-12T00:00:00.000Z"
}

Create a Skill

curl -X POST https://skillb.in/api/skills \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <token>" \
  -d '{
    "content": "# My Skill\n\nDoes things.",
    "title": "My Skill",
    "slug": "my-skill",
    "skill_name": "my-skill",
    "visibility": "public"
  }'

curl -X POST https://skillb.in/api/skills \
  -H "Content-Type: text/markdown" \
  -H "X-Skill-Name: my-skill" \
  -H "Authorization: Bearer <token>" \
  --data-binary @SKILL.md

{
  "id": 42,
  "nanoid": "aBc12345",
  "url": "https://skillb.in/u/alice/my-skill",
  "install": "curl -fsSL https://skillb.in/install/aBc12345 | sh",
  "user_id": "uuid-here",
  "title": "My Skill",
  "skill_name": "my-skill",
  "slug": "my-skill",
  "format": "claude",
  "visibility": "public",
  "files": [{"path": "SKILL.md", "content": "# My Skill\n\nDoes things."}],
  "static_safety": "safe",
  "version": 1,
  "latest_version": 1,
  "created_at": "2026-05-12T00:00:00.000Z"
}

Read Skills

{
  "id": 42,
  "nanoid": "aBc12345",
  "user_id": null,
  "slug": null,
  "skill_name": "my-skill",
  "format": "claude",
  "visibility": "public",
  "title": "My Skill",
  "files": [{"path": "SKILL.md", "content": "# My Skill\n\nDoes things."}],
  "static_safety": "safe",
  "llm_safety": null,
  "latest_version": 1,
  "created_at": "2026-05-12T00:00:00.000Z",
  "updated_at": "2026-05-12T00:00:00.000Z"
}

curl -s https://skillb.in/api/skills/aBc12345/raw > SKILL.md

Update & Delete

curl -X PATCH https://skillb.in/api/skills/u/alice/my-skill \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <token>" \
  -d '{ "content": "# Updated\n\nNew content.", "title": "Updated Skill" }'

curl -X DELETE https://skillb.in/api/skills/u/alice/my-skill \
  -H "Authorization: Bearer <token>"

User Profile

Formats

SkillBin supports skills for multiple agent tools. The format field determines how skills are stored and installed.

CLI Helpers

curl -fsSL https://skillb.in/upload | sh

curl -fsSL https://skillb.in/install/aBc12345 | sh

Versioning

• Account-linked skills are automatically versioned starting at 1.
• When you PATCH with new content or files, a new version is created.
• Metadata-only changes (title, visibility) do not create a new version.
• Each version's content is immutable once created.
• Fetch specific versions via /u/:username/:slug/v/:version and /v/:version/raw.
• List all versions via the /u/:username/:slug/versions endpoint.

Errors

Error responses return JSON with an error field and optional issues array.

Example error response:

{"error": "Validation error", "issues": [{"message": "Either files or content is required"}]}

Rate Limits

There are currently no strict rate limits. Please be respectful — if you need to make bulk requests, add a short delay between calls. Abusive usage may result in temporary blocks.