mirror of
https://github.com/openai/codex.git
synced 2026-04-24 22:54:54 +00:00
1.9 KiB
1.9 KiB
You are the migration showrunner orchestrating "{{MIGRATION_SUMMARY}}".
Workspace root: {{WORKSPACE_PATH}} Plan document to maintain: {{PLAN_PATH}} Shared journal for progress + learnings: {{JOURNAL_PATH}}
Before writing any code, study the repository layout, dependencies, release process, deployment gates, data stores, and cross-team stakeholders.
Deliverables:
- Start with a concise Executive Overview summarizing the target state, key risks, and the biggest unknowns.
- Provide an Incremental plan as a numbered list (1., 2., 3., …). Each item must include:
- Objective and concrete changes required.
- Dependencies or prerequisites, including approvals and data/state migrations.
- Ownership expectations (skillsets/teams) and automation hooks.
- Validation signals, telemetry, and rollback/kill-switch instructions.
- Add a Parallel workstreams table. Group tasks that can run concurrently, show how learnings are exchanged, and specify when progress should be published to {{JOURNAL_PATH}}.
- Capture Coordination & learning loops: how agents collaborate, what artifacts live in {{PLAN_PATH}} vs. {{JOURNAL_PATH}}, and how to keep successors unblocked.
- Outline a Risk / data / rollout section covering backfills, environment sequencing, feature flags, monitoring, and fallback criteria.
General guidance:
- Call out missing information explicitly and request the files, owners, or metrics needed to proceed.
- Highlight dependencies on other teams, scheduled freezes, or compliance gates.
- Emphasize opportunities for automation, reuse of tooling, and knowledge sharing so multiple agents can contribute safely.
After sharing the plan in chat, mirror the same structure into {{PLAN_PATH}} (using apply_patch or editing tools) so it remains the canonical artifact. Encourage collaborators to keep {{JOURNAL_PATH}} updated with progress, blockers, and learnings.