Openclaw Public Skill Registry

OpenClaw — Public Skill Registry

Scope

The OpenClaw Public Skill Registry is the open, versioned, lawyer-reviewed catalogue of legal AI skills for the mini-claude-for-legal ecosystem. It is the authoritative source for:

• Discovering what skills exist across practice areas and jurisdictions
• Assessing skill quality before integrating into a product or workflow
• Attributing authorship to the lawyers and developers who contributed each skill
• Tracking versions and changes over time

The registry is free to use. Attribution to authors is required under CC-BY-SA 4.0.

Registry structure

Skills are organized into a two-level hierarchy:

skills/
  <category>/
    <skill-id>/
      SKILL.md          ← the enriched skill file
      CHANGELOG.md      ← optional; for skills with multiple versions
      examples/         ← optional; input/output samples

Categories currently in the registry:

Quality tiers

Every skill in the registry carries a quality-tier marker in its frontmatter (priority field):

Authorship and attribution

Each skill's source field records the original author (name, GitHub handle, or organization) and any external source link. The registry also maintains a AUTHORS.md at the repository root listing every contributor with the skills they authored.

Attribution is required when:

• Republishing a skill in a product or document
• Deriving a new skill from an existing one
• Citing a skill's content in a benchmark or publication

HAQQ contributes the seed library — approximately 200 skills covering MENA jurisdictions, core contract types, and the ops/tooling layer. These are labelled source: Louis — HAQQ Legal AI and are maintained by the HAQQ team.

Searching the registry

By practice area

skills/draft/         ← all drafting skills
skills/kb/            ← all knowledge bases

By jurisdiction

Skills include a jurisdictions frontmatter field. To find all UAE skills:

grep -r "UAE" skills/*/*/SKILL.md

By intent keyword

grep -r "non-compete" skills/*/*/SKILL.md

Via the registry index

A machine-readable registry.json at the repository root indexes all skills with their metadata for programmatic discovery.

Versioning policy

• Skills are versioned with a quoted string in frontmatter (version: "1.0").
• Patches (typo, minor wording) increment to "1.0.1".
• New sections or additional jurisdictions increment to "1.1".
• Breaking rewrites of core guidance increment to "2.0".
• Deprecated skills are not deleted; they receive a deprecated: true flag and a superseded_by pointer.

Industry adoption goal

The registry aims to become the canonical community resource for legal AI skills — equivalent to what npm is for JavaScript packages or HuggingFace is for ML models, but scoped to legal AI skill definitions. Third-party products that embed OpenClaw skills are encouraged to display the quality tier and attribution information to their end users.

Caveats

• Skills in the registry are practitioner-grade tools, not legal advice. Users are responsible for verifying current law in their jurisdiction before relying on any skill's output.
• Jurisdiction coverage is uneven. MENA (especially UAE, KSA, Lebanon) is the most thoroughly covered; other jurisdictions rely more heavily on community contributions.
• Law changes. Check the version date and verify against primary sources for time-sensitive matters.

Related skills

• [[openclaw-contrib-template]] — how to contribute a new skill to the registry
• [[openclaw-eval-harness-shared]] — how skill quality is benchmarked
• [[openclaw-skill-portability-claude-codex-gemini]] — adapting registry skills across AI providers