Codex orchestration

Codex Orchestration

This skill defines the repo-local orchestration workflow for sc-lint.

Model

• team-lead coordinates sprint sequencing, worktree assignments, and PR flow
• clint is the sole developer for Codex-driven implementation work
• quality-mgr runs the QA gate after each delivery

Preconditions

Before starting a sprint:

1. docs/requirements.md, docs/architecture.md, and docs/project-plan.md define the sprint or phase review target.
2. A worktree exists for the sprint branch under the repo’s worktree strategy.
3. The target branch for the sprint is chosen from the current repo plan.
4. The following prompts exist in .claude/agents/:
   • quality-mgr.md
   • req-qa.md
   • arch-qa.md
   • flaky-test-qa.md
   • installed Rust reviewers from sc-rust
5. The following QA reporting skill exists in .claude/skills/:
   • quality-management-gh/
6. quality-mgr must read:
   • .claude/assets/sc-rust/quality-mgr/quality-mgr.rust.md
7. quality-mgr must also read:
   • .claude/skills/quality-management-gh/SKILL.md
8. sc-compose is available for rendering the JSON and markdown templates.

Sprint Flow

1. team-lead assigns development to clint using dev-template.xml.j2.
2. clint ACKs, implements, commits, pushes, and reports branch plus SHA.
3. Before QA-1, clint performs a self-directed Rust best-practices sweep on the integration branch using the same review_targets planned for QA-1 and fixes all RBP findings found there. This is a developer cleanup step, not a QA surprise.
4. team-lead opens or updates the PR.
5. team-lead assigns QA to quality-mgr using qa-template.xml.j2.
6. quality-mgr launches the reviewer set:
   • req-qa
   • arch-qa
   • rust-qa-agent
   • rust-best-practices-agent in QA-1 only
   • rust-service-hardening-agent when service-runtime review is explicitly in scope
   • flaky-test-qa when test instability risk is present
7. QA-2 and later rounds must omit rust-best-practices-agent. Any unresolved QA-1 RBP findings that are not fixed in the first fix round carry to the next phase backlog automatically and are not re-raised in later QA rounds on the same sprint branch. QA-2+ stays in targeted-fix mode: use triage_records, changed_files, and carry_forward_findings_json rather than broad re-review by default.
8. If QA passes and CI is green, merge may proceed.
9. If QA fails, team-lead first runs /triaging-findings to correlate the findings across worktrees and determine the promoted fix branch. Do not send raw QA findings directly to clint.
10. After triage completes, team-lead routes concrete fixes back to clint using fix-assignment.xml.j2. For follow-up QA rounds, the routing layer must also preserve:

• triage records from qa-triage
• carry_forward_findings_json rendered by scripts/triage_carry_forward.py
• TODO findings produced by scripts/find_todos.py

Phase-End Review

For extraction-readiness or phase-close reviews, use review-template.xml.j2 to assign a read-only review to clint.

CI

Use standard GitHub CLI:

• gh pr checks <PR> --watch
• gh pr view <PR> --json mergeStateStatus,reviewDecision

Do not assume ATM-specific PR monitoring commands exist.

Assignment Templates

Use the templates in this skill directory:

• dev-template.xml.j2
• fix-assignment.xml.j2
• qa-template.xml.j2
• review-template.xml.j2
• req-qa-assignment.json.j2
• arch-qa-assignment.json.j2
• flaky-test-qa-assignment.json.j2
• reporting templates under .claude/skills/quality-management-gh/

Use the Rust assignment templates from:

• .claude/assets/sc-rust/quality-mgr/templates/

Authoritative B.4 QA-routing surfaces:

• .claude/agents/quality-mgr.md
• .claude/agents/qa-triage.md
• .claude/skills/triaging-findings/SKILL.md
• .claude/skills/todo-triage/SKILL.md
• .claude/skills/codex-orchestration/qa-template.xml.j2
• .claude/skills/codex-orchestration/fix-assignment.xml.j2
• scripts/find_todos.py
• scripts/triage_carry_forward.py

Required Message Sequence

Every ATM task message must follow:

1. ACK
2. Work
3. Completion summary
4. Completion ACK by receiver