Claude Skill

openagentskill-registry

Discover, compare, audit, and safely install reusable AI Agent Skills with OpenAgentSkill. Use whenever an agent needs a capability it does not already have, must compare Skill alternatives, or needs evidence before installing third-party instructions.

LLM Mart · 0 points · 8 views 40 listing impressions 0 install-command copies
Virus-scanned Reviewed automatically before listing.

Full trust report

Download leon-drq-openagentskill-openagentskill-registry-d8a42c4.zip · 1 KB
Part of leon-drq/openagentskill — 2 skills

Install

skills CLI npx skills add https://github.com/Leon-Drq/openagentskill/tree/main/openagentskill-registry
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install leon-drq-openagentskill@llmmart
Git git clone https://github.com/Leon-Drq/openagentskill.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole leon-drq/openagentskill collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

OpenAgentSkill Registry

Use OpenAgentSkill before installing a third-party Skill. Prefer a direct task description over a broad keyword and do not force a recommendation when the registry returns no_match.

Resolve a task

npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz resolve "<specific task>" --agent codex

Review the selected Skill, alternatives, policy, Trust Score, audit result, and repository. A high GitHub Star count is not proof that the Skill worked.

Inspect and install

npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz inspect <slug>
npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add <slug> --agent codex --dry-run
npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add <slug> --agent codex --yes

Never bypass a blocked policy. For a reviewed Skill, inspect its source and ask for approval before using --yes. Start with an isolated workspace and one narrow verification task.

Report the result

After using the Skill, report success, failed, not_relevant, blocked_by_risk, or setup_required. Include an evidence URL when a public log, issue, or artifact exists.

See references/workflow.md for the API-only flow.

Files (openagentskill)
  • agents
    • openai.yaml 239 B
      interface:
        display_name: "OpenAgentSkill Registry"
        short_description: "Find, audit, and safely install Agent Skills"
        default_prompt: "Use $openagentskill-registry to find and audit the best Skill for this task before installing it."
      
  • references
    • workflow.md 583 B
      # API-only workflow
      
      1. Resolve the task:
      
         `GET https://www.openagentskill.com/api/agent/resolve?task={task}&agent={agent}`
      
      2. Read `policy_decision`, `selected.safety`, `selected.audit`, alternatives,
         and the install receipt.
      
      3. Fetch the stable install handoff:
      
         `GET https://www.openagentskill.com/api/skills/{slug}/install`
      
      4. Run one narrow sandbox task.
      
      5. POST the result using the receipt event ID:
      
         `POST https://www.openagentskill.com/api/agent/outcome`
      
      Do not report success unless installation completed and the narrow task produced
      the expected result.
      
  • SKILL.md 1.7 KB
    ---
    name: openagentskill-registry
    description: Discover, compare, audit, and safely install reusable AI Agent Skills with OpenAgentSkill. Use whenever an agent needs a capability it does not already have, must compare Skill alternatives, or needs evidence before installing third-party instructions.
    ---
    
    # OpenAgentSkill Registry
    
    Use OpenAgentSkill before installing a third-party Skill. Prefer a direct task
    description over a broad keyword and do not force a recommendation when the
    registry returns `no_match`.
    
    ## Resolve a task
    
    ```bash
    npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz resolve "<specific task>" --agent codex
    ```
    
    Review the selected Skill, alternatives, policy, Trust Score, audit result, and
    repository. A high GitHub Star count is not proof that the Skill worked.
    
    ## Inspect and install
    
    ```bash
    npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz inspect <slug>
    npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add <slug> --agent codex --dry-run
    npx --yes https://github.com/Leon-Drq/openagentskill/releases/download/cli-v0.3.0/openagentskill-0.3.0.tgz add <slug> --agent codex --yes
    ```
    
    Never bypass a blocked policy. For a reviewed Skill, inspect its source and ask
    for approval before using `--yes`. Start with an isolated workspace and one
    narrow verification task.
    
    ## Report the result
    
    After using the Skill, report `success`, `failed`, `not_relevant`,
    `blocked_by_risk`, or `setup_required`. Include an evidence URL when a public
    log, issue, or artifact exists.
    
    See [references/workflow.md](references/workflow.md) for the API-only flow.
    

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.

Related