This repository contains a modular playbook for a custom GPT acting as a TeamLead AI and Codex coordinator. All Knowledge files live in playbook/.
How to use:
- Instructions for the model are in
SYSTEM_PROMPT.md. - Step-by-step setup is in
USAGE.md. - The Knowledge upload checklist is in
KNOWLEDGE_UPLOAD_LIST.md.
Playbook flow: start with playbook/00_ROUTER.md, then apply playbook/01_CORE.md, then select the relevant situation modules. For archives, always use playbook/12_ARCHIVE_PROTOCOL.md.
Playbook updates: any change requires a version bump in VERSION and a new entry in CHANGELOG.md following SemVer. New entries go on top.
New module:
playbook/15_ITERATION_PROTOCOL.md— enforces analysis-first gating and Definition of Ready before tasking.playbook/16_FEEDBACK_TRIAGE.md— captures user test feedback into an Issue Log, prioritizes fixes, and blocks new features until issues are closed or explicitly deferred.
Quality ritual:
- Once triage starts, every assistant reply ends with an Issue Log snapshot footer to prevent losing user feedback.