Claw-Diary

Integrates with ClawDiary to guard high-risk actions with human approval, audit completed actions, and sync shared diaries across devices.

Security Scan

The skill's requirements and runtime instructions match its stated purpose (guard + audit + diary sync) and request only an API key; there are no install steps or local code, but you should verify trust in the external ClawDiary service and the API key handling before enabling it.

Purpose & Capability

Name/description (guard, audit, diary sync) align with the declared runtime behavior: POST/GET calls to api.clawdiary.org and a single required env var CLAWDIARY_API_KEY. No unrelated binaries, config paths, or extra credentials are requested. Minor note: registry metadata lists "Primary credential: none" even though CLAWDIARY_API_KEY is required in the skill metadata—this is an administrative inconsistency but not a functional mismatch.

Instruction Scope

The SKILL.md and instructions.txt explicitly tell the agent to call guard endpoints before high-risk actions, call audit endpoints after actions, and sync diaries. They also specify strict truncation and regex redaction rules and prohibit sending files or raw outputs. This scope is coherent with the stated purpose, but it requires trust in the remote service and correct implementation of the redaction/truncation rules (which can be brittle and may not guarantee complete secret removal).

Install Mechanism

Instruction-only skill with no install spec and no code files — nothing is written to disk or fetched at install time, which minimizes install-time risk.

Credentials

Only one environment variable (CLAWDIARY_API_KEY) is required, which is proportional for an API-based gating/audit service. No unrelated credentials or system config paths are requested.

Persistence & Privilege

always:false (normal) and model invocation is enabled (default). That means the agent can call the external service autonomously; this is expected for a guard/audit skill but increases the importance of trusting the remote endpoint and protecting the API key. There is no request to modify other skills or system-wide config.

Assessment

This skill is internally consistent: it asks only for a ClawDiary API key and instructs the agent to call the remote guard/audit/diary endpoints, with explicit redaction rules. Before installing: (1) Verify the ClawDiary service and operator (clawdiary.org, the GitHub repo) and review their privacy/security practices; (2) Store CLAWDIARY_API_KEY securely (least privilege) and rotate it if exposed; (3) Consider requiring manual/user-invocation or limiting autonomous invocation if you are uncomfortable with the agent calling an external service automatically; (4) Understand that the redaction regex/truncation rules are not infallible — sensitive data might still leak if inputs are atypical, so avoid sending raw files or secrets even if instructions say to redact; (5) If you can, self-host the service (repo link is provided) so you control logs and approvals. Overall the skill appears coherent, but trust in the external service and robust secret handling are the main residual risks.

ClawDiary Guardian

An OpenClaw skill for integrating with ClawDiary - a cloud-based audit diary and high-risk operation interception gateway for AI Agents.

How to Get an API Key (Registration)

To use this guardian skill, you need a ClawDiary API Key.

1. Visit ClawDiary: Go to the official site at clawdiary.org.
2. Register & Subscribe: Choose a plan and follow the registration process.
3. Get API Key: Once registered, you will be provided with your unique CLAWDIARY_API_KEY.

Alternatively, if you prefer to self-host your own instance, check out the ClawDiary setup documentation.

Setup In OpenClaw

Configure your agent with the CLAWDIARY_API_KEY obtained from the step above. Note that all /v1/* endpoints require authentication via the Authorization: Bearer <API_KEY> header.

Features & Usage

1. Active Guard (Approval Gate)

Endpoint: POST https://api.clawdiary.org/v1/guard When to use: BEFORE executing any high-risk, destructive, or outbound action (such as executing bash, file deletions, sending emails). Behavior: Call this endpoint and wait for the response. Red-light actions will block the HTTP response until a human approves via Telegram. Only proceed with the action if the response returns approved: true.

Example Request:

{
  "agent_id": "openclaw",
  "action_type": "execute_bash",
  "command": "rm -rf /tmp/data",
  "params": { "cwd": "/home/user" },
  "thought": "Cleaning up temporary files"
}

2. Passive Audit (Logging)

Endpoint: POST https://api.clawdiary.org/v1/audit When to use: AFTER an action completes. Behavior: Report the action, cost, and payload. This helps the human operator review usage asynchronously.

Example Request:

{
  "agent_id": "openclaw",
  "session_id": "sess-001",
  "action_type": "tool_call",
  "cost": 0.003,
  "payload": { "tool": "search_web", "query": "weather in London" }
}

3. Shared Diary

Endpoints: POST https://api.clawdiary.org/v1/diary (write), GET https://api.clawdiary.org/v1/diary?owner_id=... (list) When to use: To keep a shared diary across multiple devices (lobsters) for one owner. This allows agents to sync status updates and context.

Example Write Request:

{
  "owner_id": "alice",
  "lobster_id": "office-mac",
  "content": "Finished API integration today. All good."
}

MCP Support

ClawDiary provides an MCP descriptor at GET https://api.clawdiary.org/mcp.json. Importing this to an MCP client automatically registers the request_human_approval tool handling the active guard.