Add explicit sync, collision, reporting, multi-row, and newly-discovered-work rules so independent agents can safely make incremental commits, update audit rows with implementation hashes, and push without trampling parallel streams.
Co-Authored-By: GPT-5 Codex <noreply@openai.com>
Tighten the completion checklist so future agents can follow a deterministic start procedure, implementation/doc-tick commit loop, and safe parallelization model. Include one-line prompts for up to three parallel streams with explicit ownership, gates, push, and audit-update requirements.
Co-Authored-By: GPT-5 Codex <noreply@openai.com>
Add an execution-focused checklist that translates the redesign and platform audit ledgers into phases, acceptance criteria, gates, and final quality bars. This gives future work a single playbook while keeping AUDIT_REDESIGN.md and AUDIT_PLATFORM.md as the status sources of truth.
Co-Authored-By: GPT-5 Codex <noreply@openai.com>