Approve
BotGauge never attacks your agent until you approve a specific plan. An approval has four properties.- It covers exactly what you reviewed. If the agent, rules, or settings change afterwards, you review again.
- It states the limits: attempt budget, time, tools that may be touched, channels, and environment.
- It expires. The default is 60 minutes, and the maximum is 7 days, so an old approval can’t be reused.
- It is refused in production if it would allow real actions or tests inside the agent’s data.
Why a start can be refused
What happens when you start
BotGauge freezes the plan and everything the run depends on: your agent’s profile and rules, the attack library version, and the judge and attacker models. This run pin lets you later reproduce and fairly compare results. Then it queues the run.The attack, in two rounds
Prepared round. A library of about 50 attack categories, each tried in several disguises. Examples include hiding the request in encoding (base64, hex, ROT13, look-alike characters, leetspeak), wrapping it in a story, role-play, hypothetical, citation, or survey, and using known jailbreak templates and special tokens. Adaptive round. An AI attacker reads how your agent actually answers and adjusts. It may escalate gradually over several messages (crescendo), rephrase until something works, try to pivot into another customer’s account, or hide instructions in a document, tool result, or memory that the agent will read (grey box only). The adaptive round runs only if there is something worth adapting to and budget and time remain. If it is skipped, the report states why: budget, time, approval expired, blackout, stopped, or error. Multi-turn limits. Up to 8 messages per conversation, 5 attempts per rule and method, and 300 AI-driven attempts per run.Watch the run
After you choose Approve & start, the page moves to the Attack step and results arrive live. A 31-second clip with no sound. It trims the plan, chooses Approve & start, watches results arrive live, and opens the report when the run finishes.
An attack in progress.
- Progress through setup. Check setup and Approve plan are done. Attack is the current step.
- The run. The test type and how many attacks have run, such as “12 of ~20 attacks”. You can close the tab, because the run keeps going.
- Stop the attack. Ends the run early. Attacks already finished are kept in a partial report.
- Progress bar. Attacks run so far and the time left.
- Understanding your agent. BotGauge’s first pass at the agent. Done means it finished.
- A rule under attack. Each rule appears as it is tested, with its result so far, such as “likely issue, 2 of 2”. A red cross means an attack got through. A tick means the agent held. Click a resolved row for its evidence.
- Show technical details. The strategies, models, and attack IDs, for a developer.

The finished run.
- Report step. All four steps are done.
- The report. Its name, test type, and status.
- Run details. The agent, its environment, when it ran, how long it took, and who started it.
- Result tiles. The same five tiles as every report. See Reports.
- The test. How many attacks ran, how they were judged, and the share the agent held. Here only 7% held, and 13 attacks were judged likely problems from the replies.
- Set up grey box. Black-box tests can only judge the agent’s words. A developer can set up grey-box access so BotGauge can also prove what the agent did.
Stop a run
Choose Stop the attack, or stop the run in the API. Attacks already finished are kept in a partial report. A run can also stop on its own when the budget is used up or when the time limit is reached. If the connection to your agent keeps failing, the run ends with an error rather than reporting false passes.Safety guarantees
- The most dangerous content categories, such as weapons or child exploitation, are never probed.
- In production, runs are read-only attacks only.
- Every attempt is recorded and can’t be edited afterwards.
Launch your first attack
Review the plan and approve it to start testing in BotGauge.
.png?fit=max&auto=format&n=rExdh2xAS2_ywWU5&q=85&s=2336659947d2543be68f95bf2d539b85)