Answer: Use the eval assertions t.calledTool("inspect_pr") to confirm the agent called the specific tool and t.notCalledTool("approve_pr") to confirm it did not call another tool. Status: Confirmed Entry: tool/cursor/agents-and-automation/assert-tool-calls Revision: 1 Digest: sha256:dd0165b92f975c04 (changes whenever the content changes) Timeline: Created: 2026-11-13T03:08:26Z Last changed: 2026-11-13T03:08:26Z Last verified: 2026-11-20T12:32:40Z Next check: 2026-11-21T12:32:40Z Check interval: Daily Expected change: No fixed schedule Data: Field | Value | Source | As of called tool assertion | t.calledTool("inspect_pr") | [1] | 2026-11-20 not called tool assertion | t.notCalledTool("approve_pr") | [1] | 2026-11-20 success assertion | t.succeeded() | [1] | 2026-11-20 reply check | t.check(t.reply, includes(/ready|blocked|approve/i)) | [1] | 2026-11-20 Sources: [1] Anysphere. "cursor.com/docs/agent-sdk/evals.md". https://cursor.com/docs/agent-sdk/evals.md Role: primary. Accessed: 2026-11-20T12:32:40Z Quote: "t.calledTool("inspect_pr");" (Supports: called tool assertion) Quote: "t.notCalledTool("approve_pr");" (Supports: not called tool assertion) Quote: "t.succeeded();" (Supports: success assertion) Quote: "t.check(t.reply, includes(/ready|blocked|approve/i));" (Supports: reply check) Snapshot SHA-256: 582c4d2c88f54719b10b46b027c5bd4b4de34bb4e6b57d84dc80bd00a1c00a5b (captured 2026-11-13T03:08:26Z) Verification: Method: Automated extraction from [1] with deepseek-flash (JSON output, thinking disabled). Every value was checked by code to appear verbatim in a quoted passage of [1]. On this check the text of [1] was unchanged since 2026-11-13T03:08:26Z, so the values verified then were kept. Cross-check: not-available at 2026-11-20T12:32:40Z. No independent second source is configured for this entry Last successful check: 2026-11-20T12:32:40Z Last attempt: 2026-11-20T12:32:40Z Notes and limitations: Extracted automatically from a single official source (section "Protect a tool decision"). Related: https://latestn.com/tool/cursor/agents-and-automation/author-agent-sdk-hook How do I author a hook for the Cursor Agent SDK? https://latestn.com/tool/cursor/agents-and-automation/automation-tools What tools can a Cursor automation use? https://latestn.com/tool/cursor/agents-and-automation/automation-triggers What triggers can I use to start a Cursor automation? https://latestn.com/tool/cursor/agents-and-automation/agent-sdk-review-triage-templates What review and triage templates does the Cursor Agent SDK provide? https://latestn.com/tool/cursor/agents-and-automation/arm64-cloud-agents Does Cursor support arm64 for Cloud Agents, and how do I run agents on arm64 hardware? https://latestn.com/tool/cursor/agents-and-automation/ask-agent-self-improve How do I ask a Cursor agent to improve itself with one change? https://latestn.com/tool/ All entries in AI tools Change log: Time | Event | Change | Source 2026-11-13T03:08:26Z | added | Entry created | — Changes for this entry as a feed: https://latestn.com/tool/cursor/agents-and-automation/assert-tool-calls/latest Challenge this entry: Open challenges: None How: If a value here is wrong or out of date, send what is wrong and a link to evidence. Challenges are queued and checked against the evidence; they do not change the entry by themselves. If the entry changes as a result, the change log says so. Form: https://latestn.com/dispute/tool/cursor/agents-and-automation/assert-tool-calls POST: https://latestn.com/dispute/tool/cursor/agents-and-automation/assert-tool-calls with claim, evidence_url, and optionally field and contact (form-encoded or JSON) GET: https://latestn.com/dispute/tool/cursor/agents-and-automation/assert-tool-calls?claim={what is wrong}&evidence_url={link} How to cite: LatestN. (2026). How do I check that a Cursor agent called a specific tool and not another in an eval? Revision 1, last verified 2026-11-20T12:32:40Z. https://latestn.com/tool/cursor/agents-and-automation/assert-tool-calls (accessed 2026-11-21T09:58:16Z). @misc{LatestN_tool_cursor_agents_and_automation_assert_tool_calls_r1, title = {How do I check that a Cursor agent called a specific tool and not another in an eval?}, author = {{LatestN}}, year = {2026}, howpublished = {\url{https://latestn.com/tool/cursor/agents-and-automation/assert-tool-calls}}, note = {Revision 1; digest sha256:dd0165b92f975c04; last verified 2026-11-20T12:32:40Z; accessed 2026-11-21T09:58:16Z} } Reuse: License: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) Attribution: Copying and adapting are allowed, including commercially, if you credit LatestN and link to the entry you used. Each entry ends with a ready-made citation. URL: https://latestn.com/tool/cursor/agents-and-automation/assert-tool-calls