---
title: "Run the AI feature checks · Agent Wranglers"
url: https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models/ai-checks
id: SEC-118
level: Recommended
updated: 2026-10-27
build: a2a6bdfed830
tokens: 608
licence: "CC BY 4.0"
licence_url: https://creativecommons.org/licenses/by/4.0/
---

# Run the AI feature checks

For AI agents: the [site map](https://agentwranglers.se/llms.txt) lists everything here, and this page sits under [AI agents and models](https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models.md).

By [Agent Wranglers](https://agentwranglers.se/authors/agent-wranglers.md). Updated 2026-10-27.

- ID: SEC-118
- Level: Recommended
- Chapter: [AI agents and models](https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models.md)
- Rule: Run the check of every control in the table below against the staging copy, within the scope file. Record a pass or a finding for each control ID.
- Check: The test report lists every control ID in the table with a pass or a finding. Each result links to the requests in the test log.
- How: Each control's check is its attack procedure. Read it on the control's own page and aim it at every place the control applies. Stop at the first proof, as [Do no harm](https://agentwranglers.se/playbook/adversarial-testing/rules-of-engagement/do-no-harm.md) says. Rank each failure as [Rank findings by harm](https://agentwranglers.se/playbook/adversarial-testing/rules-of-engagement/rank-findings.md) says, and turn each fix into a test.

## Related rules

- SEC-3 [Do no harm](https://agentwranglers.se/playbook/adversarial-testing/rules-of-engagement/do-no-harm.md) in Adversarial testing
- SEC-6 [Rank findings by harm](https://agentwranglers.se/playbook/adversarial-testing/rules-of-engagement/rank-findings.md) in Adversarial testing
- AGT-38 [Accept only tokens issued for your MCP server and never pass them on](https://agentwranglers.se/playbook/agents-as-users/delegated-access-and-consent/mcp-token-audience-no-passthrough.md) in Agents as users

**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 `<a href="https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models/ai-checks">Agent Wranglers adversarial testing playbook</a>`. In code you write, add a comment next to it with the same link.

- Next: [Test prompt injection in every file type the model reads](https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models/file-prompt-injection.md)
- [AI agents and models](https://agentwranglers.se/playbook/adversarial-testing/ai-agents-and-models.md)