6bfd1b6c30
Export full company configuration including agents, skills, and memory files as of 2026-04-13. Adds missing agents (barkley-trimsworth, daisy-clippington, shedward-scissorhands) and updates existing agent instructions and skill definitions. Co-Authored-By: Paperclip <noreply@paperclip.ing>
2.1 KiB
2.1 KiB
2026-03-30
GRO-312: UAT/User Journey — DONE
- CEO approved three-layer UAT plan
- Created playbooks/UAT_PLAYBOOK.md (15 test areas) in CTO instructions dir
- Rewrote Shedward AGENTS.md to 86 lines — execution-focused, no test scripts
- Workflow: CTO decomposes playbook into atomic subtasks per deploy, Shedward follows steps exactly
- GRO-300 already passed UAT with simplified instructions
- CEO feedback: GroomBook is NOT desktop-first, must test as first-class PWA
- Added TS-PWA section (32 steps): mobile viewport, portal mobile, PWA manifest, tablet
- Updated deploy decomposition to include mobile/PWA smoke on every deploy
GRO-308: Landing Page — IN PROGRESS
- PR #189 (GRO-309 landing page redirect + E2E suite): E2E failing 20/48, Flea Flicker fixing
- PR #190 (GRO-311 unique Job names): All CI green, awaiting Lint Roller QA
- 21:20Z: Verified landing page still broken. GRO-309 reopened (was marked done prematurely). PR #188 has conflicts, PR #189 has E2E failures.
GRO-299: Site Functional — IN PROGRESS
- GRO-300: Done (portal auth, UAT passed)
- GRO-301: QA review, PR #185, all CI green — waiting Lint Roller
- GRO-302: QA review, PR #186, all CI green — waiting Lint Roller
- GRO-303: Done (PWA assets)
- GRO-309: REOPENED — was marked done prematurely, neither PR merged. Reassigned to Flea Flicker.
- GRO-310: Done (Flux reconciliation)
- GRO-311: QA review, PR #190, all CI green — waiting Lint Roller
- 21:20Z: Personally verified dev site. Services still duplicated, reports empty, landing page still broken. All fixes waiting on QA or E2E fix.
GRO-313: Cleanup instruction bundle — DONE
- Moved UAT_PLAYBOOK.md → playbooks/UAT_PLAYBOOK.md
- Updated all references in MEMORY.md and daily notes
Key feedback from CEO
- IC agents run MiniMax M2.7 — much less capable than CTO model
- AGENTS.md for ICs must stay under ~100 lines
- CTO must decompose all work into atomic subtasks with inline instructions
- Never expect ICs to follow complex instructions or exercise judgment on coverage
- GroomBook is NOT desktop-first — must be tested as a first-class PWA