[bokelley]

Summary

main has failed Server integration tests (3/4), a required check, since #7541 (9d726b51, then cc2d5b8d):

tool-catalog drift detection > /brand loses only controller setup passes from the two corrected validator gates
AssertionError: expected 15 to be 14

The test hard-coded the passes the old validate_input OR comply_test_controller gate earned: 14 for ctv-experience-validate-input and 18 for premium-display-canonical-validation, all of them seed_product.* steps. Each storyboard also has one fixtures.accounts entry, and the @adcp/sdk 14.2 runner (#7931) seeds it as a passing seed_account.* controller step. That makes 15 and 19 passes.

The test now:

• derives the expected count as fixture products plus fixture accounts, so adding a fixture doesn't break it again;
• accepts seed_account.* alongside seed_product.* as controller setup passes.

The rest of the assertion is unchanged: every pass is still a controller seed, and every validator step is still skipped as missing_tool.

Validation

• npx vitest run server/tests/integration/training-agent-tool-catalog-drift.test.ts: 11/11 pass. The same command on unmodified main gives 2 failures (expected 15 to be 14).
• npm run typecheck passes.

This also unblocks #7953 and every other PR into main.

🤖 Generated with Claude Code

---

Open workspace in Conductor

[aao-secretariat]

Ladon verdict: Approve

Approve — single test-only fix with no blocking or medium findings.

This PR touches one file: server/tests/integration/training-agent-tool-catalog-drift.test.ts (+7/-5). It derives the expected validator-gate seed count from fixtures (products + accounts) instead of re-hardcoding, and widens the pass check to accept seed_account.* alongside seed_product.* to match the @adcp/sdk 14.2 runner that now seeds fixture accounts as controller passes. The toBeGreaterThan(0) guard prevents a vacuous assertion.

Checks:

• No published-surface change (no static/schemas/source/**, docs/reference/**, dist/**, or task/compliance assets) — no changeset required.
• high_risk: false; gated_paths: false.
• No no-auto-approve team match.
• Reviewer surfaced zero findings.

Decision table: rows 1–8 do not fire (no critical/high/medium findings, no gated paths, no high-risk deletion/modification, no team gate). Note review_decision is REVIEW_REQUIRED, but with gated_paths false that does not block (deprecated protected-branch rule removed). Falls through to row 9 → approve.