Tool Policies
Agents in Selu can use tools — things like searching the web, reading files, calling APIs, or sending messages. Tool policies give you fine-grained control over which tools each agent is allowed to use, and whether you’d like to approve certain actions before they happen.
How tool policies work
Section titled “How tool policies work”Every tool an agent has access to can be assigned one of three policies:
The agent can use this tool freely, without asking. This is the default for most low-risk tools like answering questions or doing basic lookups.
The agent is blocked from using this tool entirely. Useful if you want to disable capabilities you’re not comfortable with.
The agent must ask for your approval each time it wants to use this tool. You’ll receive a notification on the active channel and can approve or deny the action.
Configuring policies
Section titled “Configuring policies”- Go to Settings → Security → Tool Policies.
- Select the agent you want to configure.
- You’ll see a list of all tools the agent has registered. For each tool, choose Allow, Deny, or Prompt.
- Save your changes.
Changes take effect immediately — no need to restart the agent.
The approval flow
Section titled “The approval flow”When a tool is set to Prompt, here’s what happens:
- The agent decides it needs to use the tool.
- Selu pauses the agent and sends you a notification: “Agent wants to use [tool name]. Reason: [agent’s explanation]. Allow?”
- You reply with yes or no (or click the approve/deny buttons in web chat).
- If approved, the agent proceeds. If denied, the agent adapts and tries a different approach.
Approval requests time out after 5 minutes by default. If you don’t respond, the action is denied.
Default policies
Section titled “Default policies”When you install a new agent, Selu applies sensible defaults:
| Tool category | Default policy |
|---|---|
| Read-only (search, fetch, lookup) | Allow |
| Communication (send email, post message) | Prompt |
| Destructive (delete, modify, write) | Prompt |
| System (execute commands, file system access) | Deny |
You can override any of these defaults per agent.
Global vs. per-agent policies
Section titled “Global vs. per-agent policies”- Global policies apply to all agents unless overridden. Set these under Settings → Security → Tool Policies → Global.
- Per-agent policies override the global setting for a specific agent. Set these from the agent’s individual policy page.
Audit log
Section titled “Audit log”Every tool usage — whether allowed, denied, or approved via prompt — is recorded in the audit log under Settings → Security → Audit Log. This gives you a full history of what your agents have done and the decisions you’ve made.