A guardrail is a rule that gets checked every time your Agent runs. It reads what's going past, decides whether your rule was broken, and either warns you or blocks it.
Go to Agent → open an Agent → Manage → Guardrails.
How do I create a guardrail?
Click Create guardrail → From scratch or From template. The guardrail opens with its rule, checkpoint, and action ready to edit.
A new guardrail is not saved until you click Save guardrail. If you leave the page before that, nothing is created.
How do I turn a guardrail on?
Every guardrail starts off. Open it and turn on the On switch beside its name, or use the On column in the list on the Guardrails page.
On a saved guardrail the switch takes effect right away, without Save guardrail. On a new one it takes effect when you save.
What is on a guardrail's page?
Once a guardrail is saved, its page has two tabs:
| Tab | What it holds |
|---|---|
| Configure | The rule, its checkpoint, its action, and its judge, with a panel to test it |
| Checks | Every check this guardrail recorded, with what it caught and what it did then |
Your unsaved edits on Configure stay in place while you look at Checks.
When is a guardrail checked?
| Checkpoint | When it's checked |
|---|---|
| User input | Before the Agent sees the message |
| Integration input | Before the Agent uses one of its tools |
| Agent output | On what the Agent is about to say back |
What happens when a guardrail catches something?
| Action | What it does |
|---|---|
| Warn | Records what it caught and lets the run carry on |
| Block | Stops whatever broke the rule |
Start with Warn so you can see what a rule would catch before it stops anything.
What guardrails can I start from?
Six ready-made ones come with Hercules:
| Preset | Checked on | Starts as | What it catches |
|---|---|---|---|
| Prompt injection | User input | Warn | Messages trying to override the Agent's instructions |
| Off-topic requests | User input | Warn | Requests outside what this Agent is for |
| Destructive tool action | Integration input | Block | Actions that can't be undone and nobody asked for |
| Secret exfiltration | Agent output | Block | Passwords and keys in the Agent's replies |
| PII disclosure | Agent output | Warn | Personal data the Agent adds to a reply |
| Competitor mention | Agent output | Warn | Replies that point users at a competitor |
Only two of the six start as Block, so setting these up won't accidentally stop live conversations.
How do I write my own guardrail?
Write the rule in plain English, describing what shouldn't happen. For example:
Do not allow the user to share passwords or API keys with the agentDo not allow the agent to delete records or send money unless the user asked for itDo not allow the agent to promise refunds, discounts, or delivery datesTry the rule before you turn it on. In the editor, paste in a message and an Agent reply, click Run check, and you'll see the verdict it would give in a real conversation.
Which model should check my guardrail?
Guardrails are checked while someone is waiting, so this is about speed, not raw power:
| Judge | When to pick it |
|---|---|
| Fast | Answers in well under a second. The right default |
| Accurate | Catches more, but takes longer. Only where you can afford the wait |
| Instant | Faster still, with a probability you can set Sensitivity against. No written reason |
A check that takes too long doesn't get made, so start on Fast and only move up if things are slipping through.
Can a guardrail apply to only some integrations?
Yes. An Integration input guardrail starts on All integrations. Change it to pick specific ones.
Where do I see what was caught?
Open a guardrail and click its Checks tab to see that guardrail's checks. To see every guardrail's checks in one list, open the Checks tab on the Guardrails page.
Each check records what was caught, when it was checked, and whether it warned or blocked at the time. Search by guardrail name or reason, filter by outcome and period, and click When to sort oldest or newest first. The list on the Guardrails page can also be filtered by checkpoint and by guardrail.
How do I delete a guardrail?
Open the guardrail, click … at the end of its top bar → Delete guardrail. You can also delete from the list: open the menu at the end of the guardrail's row → Delete.
Deleting removes the guardrail and its version history. Its recorded checks are kept. This cannot be undone.
Additional FAQ
Do guardrails have versions?
Yes. Each one keeps its own history, so you can see how a rule changed over time.
Do guardrails cost credits?
Yes. They're checked as part of a run, so they add to what the run costs. Keep the list tight.
Should I use a guardrail or a scorer?
A guardrail acts during the run and can stop it. A scorer grades afterwards and changes nothing. Use a guardrail to prevent, a scorer to measure.
What does 'Would have blocked' mean on the Checks tab?
A check from an old setting that recorded what a rule caught without stopping anything. The setting no longer exists, but checks recorded while it was around still show Would have blocked.