safe-action

# safe-action

Measure twice, cut once. Before taking destructive, irreversible, or high-stakes actions, this skill runs a three-part safety pre-flight: risk assessment, reversibility check, and timing awareness.

Combines three AgentUtil services — think (safety checklists), undo (reversibility intelligence), and context (situational timing) — into a single decision workflow.

## When to Activate

Use this skill before:

- **Deleting** anything — repositories, databases, user accounts, files, branches, DNS records
- **Deploying** to production — code releases, infrastructure changes, migrations
- **Mass operations** — bulk updates, batch deletes, sending to many recipients
- **Permission changes** — revoking access, changing roles, modifying security settings
- **Infrastructure changes** — scaling down, terminating instances, modifying load balancers
- **Financial actions** — processing refunds, transferring funds, modifying billing

**Do NOT use for:** read-only operations, local development changes, or actions the user has explicitly marked as low-risk.

## Workflow

### Step 1: Safety checklist (think)

Get a pre-flight checklist for the action type.

**MCP (preferred):** If `@agentutil/think-mcp` is available:
```
think_check({ action: "database_migration", description: "Drop column users.email in production" })
```

**HTTP fallback:**
```bash
curl -X POST https://think.agentutil.net/v1/check \
  -H "Content-Type: application/json" \
  -d '{"action": "database_migration", "description": "Drop column users.email in production"}'
```

Note the `risk_level` and any `critical` severity items in the checklist.

### Step 2: Reversibility check (undo)

Assess whether the action can be undone and find safer alternatives.

**MCP (preferred):** If `@agentutil/undo-mcp` is available:
```
undo_check({ platform: "GitHub", resource: "repository", action: "delete" })
```

**HTTP fallback:**
```bash
curl -X POST https://undo.agentutil.net/v1/check \
  -H "Content-Type: applica...