# How do I verify OpenClaw actions with Provely?

> Connect OpenClaw to Provely through the CLI, the REST API. The guidance files, the hooks, and the CI gate. The profile was verified on 2026-09-05.

Canonical: https://provely.sh/agents/openclaw  
Last reviewed: 2026-09-05  

**OpenClaw connects to Provely through the CLI, the REST API. The agent does the action. Provely verifies the outcome and signs a receipt. OpenClaw never holds the verifier credentials.**

A person verified the OpenClaw capabilities on 2026-09-05. Agent products change monthly. Re-verify before you build.

Source: https://provely.sh/agents (retrieved 2026-09-05)

## Which surfaces does OpenClaw use?

| Surface | Support | Detail |
| --- | --- | --- |
| [MCP server](/docs/mcp) | Verify at implementation | transport stdio |
| [CLI](/docs/cli) | Supported | Runs through the shell tool of the agent. |
| [SDK](/docs/sdk) | Not supported |  |
| [REST API](/docs/api) | Supported | Any HTTP call reaches the control plane. |

## Which guidance files does OpenClaw read?

The renderer writes Agent Skills SKILL.md from one source. A new agent needs a profile, not core code.

| Format | Path | Mode |
| --- | --- | --- |
| Agent Skills SKILL.md | `skills/provely-outcome-verification/SKILL.md` | write |

## How is a false completion claim blocked?

| Enforcement | Support | Detail |
| --- | --- | --- |
| Advisory guidance | Supported | The instruction files state the rules for every agent. |
| Lifecycle hooks | Not supported | No hook events stated. |
| CI gate | Not stated | `provely verify --wait` fails the pipeline unless the verdict is VERIFIED. |

## How do I connect OpenClaw?

### Connect OpenClaw to Provely

1. **Install the CLI.** Run `npx provely --version`, or download the static binary.
2. **Check the MCP support of the host.** The profile marks MCP as verify at implementation. When the host runs MCP, register `provely mcp`. If not, use the CLI through the shell tool.
3. **Write the guidance files.** Run `provely init --agent openclaw`. It writes `skills/provely-outcome-verification/SKILL.md` (write). The guidance carries one rule: a successful tool call is not completion.
4. **Report the verdict exactly as returned.** VERIFIED: "The action is verified complete. Receipt: <id>." PENDING: "The action is accepted but not yet verified. Operation: <id>."

## What else does the profile state?

- Verified 2026-09-05: OpenClaw follows the AgentSkills spec and reads <workspace>/skills, <workspace>/.agents/skills, and ~/.agents/skills. SKILL.md needs name and description. Source: https://github.com/openclaw/openclaw (docs/tools/skills.md).
- Checked 2026-09-05: docs/gateway/configuration.md names an `mcp` config section and `mcp.apps`. The tool-server semantics were not readable. Keep MCP at verify_at_implementation. The CLI surface works through the shell tool.
- Checked 2026-09-05: OpenClaw hooks are inbound webhooks, not lifecycle hooks. No hook enforcement exists. OpenClaw has cron jobs: schedule `provely status <operation_id>` to re-check PENDING operations.
- OpenClaw runs messaging, purchases, and device control. It is not a repository agent, so the CI gate and the scanner do not apply.

> OpenClaw keeps its action credentials, and no surface returns the verifier credentials to it. Read [the credential boundary](/docs/security). An operation outlives the session that opened it. Read [the API reference](/docs/api).

### Which host versions does the profile cover?

`>=2026.9.0`. The integrations team last verified the profile with 2026.9.1 on 2026-09-05 (npm openclaw; github.com/openclaw/openclaw package.json).

## Read next

- [Verify Claude Code actions](https://provely.sh/agents/claude-code)
- [Verify Codex actions](https://provely.sh/agents/codex)
- [Verify Resend actions](https://provely.sh/verify/resend)
- [See every agent integration](https://provely.sh/agents)
- [Read the MCP server reference](https://provely.sh/docs/mcp)
- [Read the credential boundary](https://provely.sh/docs/security)