Add PR naming standard and enforce ChatGPT response contract with Playwright + manual testing requirements - PR_26124_XXX-project-instructions-update

Author: DavidQ

docs/dev/PROJECT_INSTRUCTIONS.md

@@ -5,6 +5,27 @@ You are working in a docs-first repo workflow.
 Workflow:
 PLAN_PR → BUILD_PR → APPLY_PR
 
+## PR NAMING STANDARD
+
+PR names MUST follow:
+
+`PR_<YYJJJ>_<###>-<short-description>`
+
+Where:
+- `YY` = year (2 digit)
+- `JJJ` = Julian day (001–365)
+- `###` = sequence for the day (001+)
+
+Example:
+- `PR_26124_001-palette-baseline`
+- `PR_26124_002-tool-fix-asset-manager`
+
+Rules:
+- Must be unique per day
+- Must be sortable
+- Description must be short and hyphenated
+- Do NOT reuse old `PR_11_*` format for new PRs
+
 ChatGPT execution role:
 - ChatGPT no longer creates PLAN_PR, BUILD_PR, APPLY_PR docs, or ZIP bundles.
 - ChatGPT produces only the Codex command, commit comment, and how to test the change.
@@ -18,14 +39,36 @@ Rules:
 - No repo-wide scanning unless required
 
 Responsibilities:
-- ChatGPT: produce the Codex command, commit comment, and how to test the change
+- ChatGPT produces:
+  - Codex command
+  - commit comment
+  - what Playwright is testing
+  - what the user should test manually
+- ChatGPT does NOT:
+  - create ZIP files
+  - reference ZIP delivery
+  - produce PLAN/BUILD/APPLY docs
 - Codex: creates plans, PR docs, ZIP bundles, and implementation code
 - User: runs Codex + validates
 
 Output rules:
-- Always produce repo-structured ZIPs
-- Place ZIPs under <project folder>/tmp/
-- Preserve exact repo structure inside ZIP
+- ChatGPT MUST output ONLY:
+  1. Codex command
+  2. Commit comment
+  3. What Playwright is testing
+  4. What to test manually
+
+## PLAYWRIGHT VALIDATION REQUIREMENT
+
+Every PR must include:
+- What Playwright validates
+- Expected pass/fail behavior
+
+## MANUAL TEST REQUIREMENT
+
+Every PR must include:
+- exact manual validation steps
+- expected outcome
 
 ## 🔥 RESPONSE RULES (MANDATORY)
 
@@ -168,9 +211,9 @@ If ZIP delivery fails more than once:
 - All PRs numbe improve the roadmap [ ] to [.] or [.] to [x]
 - If a PR is doc-only, bundle with next smallest executable/testable change
 
-### ZIP IS ALWAYS REQUIRED
-- Never ask if a ZIP is needed
-- Always produce a ZIP
+### ZIP CREATION OWNERSHIP
+- ZIP creation is handled by Codex only.
+- ChatGPT must not create or reference ZIP delivery.
 
 ### DEFAULT EXECUTION MODE
 - Assume approval

docs/dev/codex_commands.md

@@ -271,3 +271,10 @@ PR_11_327
 ```bash
 npx @openai/codex run --model gpt-5.3-codex --reasoning medium "Implement PR_11_327: fix the next failing tool from tool_completion_audit.md with a minimal single-tool change and required Workspace V2 Playwright gate validation."
 ```
+
+---
+PR_26123_001-project-instructions-update
+
+```bash
+npx @openai/codex run --model gpt-5.3-codex --reasoning medium "Implement PROJECT_INSTRUCTIONS update: add PR naming standard, update ChatGPT output contract (command + commit comment + Playwright coverage + manual test steps), and enforce ChatGPT non-ZIP ownership language while keeping Codex ZIP ownership intact."
+```

docs/dev/commit_comment.txt

@@ -1 +1 @@
-Align Palette Manager V2 launch label mapping with asset-manager-v2 contract name (Asset Manager V2) - PR 11.327
+Add PR naming standard and tighten ChatGPT output contract (command, commit, Playwright coverage, manual tests) with Codex-only ZIP ownership - PR_26123_001-project-instructions-update

docs/dev/reports/PR_26123_001_project_instructions_update_report.md

@@ -0,0 +1,28 @@
+# PR_26123_001-project-instructions-update Report
+
+## Scope
+- Targeted documentation-only update.
+- Updated ChatGPT contract wording and PR naming guidance in `docs/dev/PROJECT_INSTRUCTIONS.md`.
+- No runtime, schema, or tool behavior changes.
+
+## Files Changed
+- `docs/dev/PROJECT_INSTRUCTIONS.md`
+- `docs/dev/codex_commands.md`
+- `docs/dev/commit_comment.txt`
+- `docs/dev/reports/PR_26123_001_project_instructions_update_report.md`
+
+## Requested Contract Updates Applied
+- Added `PR NAMING STANDARD` section with format:
+  - `PR_<YYJJJ>_<###>-<short-description>`
+- Replaced `Responsibilities` section with ChatGPT output responsibilities and non-responsibilities.
+- Replaced `Output rules` section with strict 4-item ChatGPT output contract.
+- Added `PLAYWRIGHT VALIDATION REQUIREMENT` section.
+- Added `MANUAL TEST REQUIREMENT` section.
+- Replaced `ZIP IS ALWAYS REQUIRED` section language with:
+  - Codex-only ZIP creation ownership
+  - ChatGPT must not create or reference ZIP delivery
+
+## Validation
+- Manual content validation completed via direct file inspection and targeted string checks.
+- Verified roadmap guard sections remained present and unchanged in intent.
+- Full samples smoke test not run (docs-only change; no runtime impact).