Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
## Summary: Security and documentation improvements; no runtime behavior changes. - Added explicit statement of execution safety guarantees (no dynamic loading, no runtime downloads, no code eval) to SKILL.md. - Enhanced SECURITY.md with clarified safety model and operator responsibilities. - Documentation files updated to emphasize static handler registry and immutable runtime. - No changes to service endpoints, environment variables, or execution logic.
What This Skill Does
OpenExec is a deterministic execution service that runs only pre-approved actions and emits a verifiable receipt for every attempt. Signature verification is fully offline with no outbound HTTP or governance calls at runtime. It enforces replay protection via nonce uniqueness and does not self-authorize, define policy, or make outbound calls during execution.
Execution, approval, and witnessing are split across three replaceable layers, so no single component can self-authorize or bypass the approval chain without network access at runtime.
When to use it
- Running approved automation tasks in air-gapped or restricted environments
- Auditing every execution attempt with tamper-evident receipts
- Enforcing a fixed allow-list of permitted actions for an AI agent
- Preventing duplicate execution of the same command via nonce tracking
- Separating execution enforcement from policy in multi-team governance workflows
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Execute the approved 'deploy' action with a unique nonce
- 1Send POST /execute with action name, payload, and nonce
- 2Check nonce uniqueness against the local database to prevent replay
- 3In ClawShield mode, verify the Ed25519 signature on the approval artifact fully offline
- 4Execute the action deterministically via the static handler registry
- 5Return the result and emit a verifiable receipt hash
Execution result plus a receipt hash, verifiable via POST /receipts/verify
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.