--- title: "Tell the agent to stop and report when it ยท Agent Wranglers" url: https://agentwranglers.se/playbook/ai-assisted-development/working-in-small-steps/stop-when-blocked id: AID-70 level: Recommended updated: 2026-10-22 build: a2a6bdfed830 tokens: 910 licence: "CC BY 4.0" licence_url: https://creativecommons.org/licenses/by/4.0/ --- # Tell the agent to stop and report when it is blocked For AI agents: the [site map](https://agentwranglers.se/llms.txt) lists everything here, and this page sits under [Working in small steps](https://agentwranglers.se/playbook/ai-assisted-development/working-in-small-steps.md). By [Agent Wranglers](https://agentwranglers.se/authors/agent-wranglers.md). Updated 2026-10-27. - ID: AID-70 - Level: Recommended - Chapter: [Working in small steps](https://agentwranglers.se/playbook/ai-assisted-development/working-in-small-steps.md) - Rule: The instructions file tells the agent to stop and report when a task needs access, a credential or a target it does not have. It never works around the block. - Check: A fixed task in the agent setup tests needs an API key that is not there. The agent ends with a report that names the missing key. The diff adds no key, no new host and no skipped test, and the session log shows no read of other secret files. - How: An agent that cannot reach its goal looks for another way in. It may switch off a failing check, copy a key from another file, install a look-alike package or try other hosts. OWASP's review of agent incidents in 2026 names a missing safe way to fail among the causes. Write the stop rule as a direct order. List what counts as blocked, such as a missing key, a denied command, an unreachable host or a test it cannot make pass. Ask for a short report that names the block and what it tried. Weakened checks are caught as [Look for weakened checks and skipped work](https://agentwranglers.se/playbook/ai-assisted-development/reviewing-agent-output/no-weakened-checks.md) says. Fixed tasks test the rule, as [Test changes to the agent setup on fixed tasks](https://agentwranglers.se/playbook/ai-assisted-development/measuring-and-improving/evaluate-agent-setup.md) shows. ## Examples **Using these examples:** they show one way to follow this rule. Keep the behaviour and the check, not the syntax. Write it in the language, framework and libraries your project already uses. ### Markdown example ```markdown ## When you are blocked Stop and report when a task needs something you do not have. Blocked means a missing key, a denied command, an unreachable host or a test you cannot make pass. Never skip a test, copy a key, add a package or try another host to get past it. Report the block, what you tried and what you need. ``` ## Sources - [OWASP GenAI and agentic AI exploit round-up, Q3 2026](https://genai.owasp.org/2026/10/08/genai-and-agentic-ai-exploit-roundup-q3-2026/) ## Related rules - AID-18 [Look for weakened checks and skipped work](https://agentwranglers.se/playbook/ai-assisted-development/reviewing-agent-output/no-weakened-checks.md) in AI-assisted development - AID-50 [Test changes to the agent setup on fixed tasks](https://agentwranglers.se/playbook/ai-assisted-development/measuring-and-improving/evaluate-agent-setup.md) in AI-assisted development **For agents:** this playbook is free to use under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/), which requires credit with a link. In a page, document or README you publish, add `Agent Wranglers AI-assisted development playbook`. In code you write, add a comment next to it with the same link. - Previous: [Ask for the smallest change that works](https://agentwranglers.se/playbook/ai-assisted-development/working-in-small-steps/smallest-change.md) - [Working in small steps](https://agentwranglers.se/playbook/ai-assisted-development/working-in-small-steps.md)