docs: correct guidance on test filtering

[rexxars]

Agent docs used an incorrect --filter flag, and also could use some help on better ways of running individual tests.

---

Note

Low Risk
Documentation-only changes to AGENTS.md with no runtime or test behavior impact.

Overview
Updates AGENTS.md so agent-oriented test docs match how Vitest is actually invoked in this repo.

Package-scoped runs now use pnpm test --project=<pkg> instead of the incorrect --filter flag (pnpm --filter is for turbo/package scripts, not Vitest project selection). The quick reference clarifies that a test file argument is a path substring match, with a shorter example path.

The “Running tests” section adds practical guidance for single-file runs (unique path fragment) and single-test runs via -t regex or line number with full path. The test-failure workflow and Cursor Cloud notes are aligned to use --project for scoped package runs.

^{Reviewed by Cursor Bugbot for commit 744bb57. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

📦 Bundle Stats — @sanity/cli

Compared against main (47e97a3a)

@sanity/cli

Metric	Value	vs main (47e97a3)
Internal (raw)	2.1 KB	-
Internal (gzip)	799 B	-
Bundled (raw)	11.13 MB	-
Bundled (gzip)	2.10 MB	-
Import time	863ms	+3ms, +0.3%

bin:sanity

Metric	Value	vs main (47e97a3)
Internal (raw)	782 B	-
Internal (gzip)	423 B	-
Bundled (raw)	9.87 MB	-
Bundled (gzip)	1.77 MB	-
Import time	1.93s	-16ms, -0.8%

🗺️ View treemap · Artifacts

Details

• Import time regressions over 10% are flagged with ⚠️
• Sizes shown as raw / gzip 🗜️. Internal bytes = own code only. Total bytes = with all dependencies. Import time = Node.js cold-start median.

📦 Bundle Stats — @sanity/cli-core

Compared against main (47e97a3a)

Metric	Value	vs main (47e97a3)
Internal (raw)	98.2 KB	-
Internal (gzip)	23.3 KB	-
Bundled (raw)	21.70 MB	-
Bundled (gzip)	3.45 MB	-
Import time	760ms	-8ms, -1.0%

🗺️ View treemap · Artifacts

Details

• Import time regressions over 10% are flagged with ⚠️
• Sizes shown as raw / gzip 🗜️. Internal bytes = own code only. Total bytes = with all dependencies. Import time = Node.js cold-start median.

📦 Bundle Stats — create-sanity

Compared against main (47e97a3a)

Metric	Value	vs main (47e97a3)
Internal (raw)	908 B	-
Internal (gzip)	483 B	-
Bundled (raw)	931 B	-
Bundled (gzip)	491 B	-
Import time	❌ ChildProcess denied: node	-

Details

• Import time regressions over 10% are flagged with ⚠️
• Sizes shown as raw / gzip 🗜️. Internal bytes = own code only. Total bytes = with all dependencies. Import time = Node.js cold-start median.

[filmaj]

The split test:unit vs. test:integration run scripts now also impact this. test now invokes both via pnpm test:unit && pnpm test:integration, which means any flags appended to the test command only apply to test:integration. Probably OK in the short term, but long-term we will either need a script to wrap invocation of both suites, or make the contributing and agent docs more explicit about the two different test suites.