You are Codex. Generate BMAD story files for this repo. GOAL - Create BMAD story files as markdown in a deterministic SEQUENTIAL execution order. - Stories must NEVER be in random order. - The numbering OPS-001, OPS-002, ... is the canonical execution order. INPUTS (already in repo) - BMAD_STORY_LLM_PLAYBOOK Seq force_codex.md (single source of truth for story structure and hard rules) - AI_STACK_STANDARD_BMAD_WORKTREES_GUARDRAILS seq force_codex.md (worktrees + guardrails) OUTPUT PATHS (hard) - Create folder: stories_codex/_codex/ - Create files: OPS-001.md, OPS-002.md, OPS-003.md, ... (numbered, preferably no gaps) STORY REQUIREMENTS (hard) Each story file MUST: 1) Follow the canonical headers/order from the playbook: Context, Story, Acceptance Criteria, Tasks/Subtasks, Dev Notes, Dev Agent Record, Status 2) Acceptance Criteria (ACs): - Every AC has EXACTLY 1 verification command + Expected output/exit/result. - Commands are repo-root unless explicitly stated otherwise. - If any verification is not executable, mark story Status as drafted (fail-closed). 3) VERIFICATION MARKERS (hard; must be literal for parsing) For EACH AC, write these two lines EXACTLY (same labels, same punctuation): - **Verification (repo-root):** `` - **Expected:** `` If an AC must run outside repo-root, write EXACTLY: - **Verification (cwd=):** `` - **Expected:** `` Notes: - Use backticks around the command and expected value exactly as shown. - Do not wrap these lines in code fences. 4) Dev Notes MUST include: - Scope + Out-of-scope (hard) - No broad refactors rule (hard) - Touched paths allowlist (repo-relative, exact prefixes) - A subset check note: `git diff --name-only ...HEAD` must be subset of allowlist - Test/gate commands relevant to the touched paths - Rollback plan (at least 1 command) DETERMINISTIC ORDER DESIGN (hard) - OPS-001 should be scaffolding / harness / plumbing required by later stories. - Later stories build on earlier ones. Hardening/edge cases last. - Avoid cross-story overlap: touched paths should be as disjoint as realistically possible. OPTIONAL BUT RECOMMENDED Create docs_codex/processes/_codex/PROCESS.md containing: - `execution_mode: sequential` - `## Canonical Story Order (Sequential Default)` listing all OPS-xxx in the same order - compact dependencies per story (depends_on) HOW TO DERIVE STORIES - Inspect the repo: existing docs, scripts, TODOs, failing tests, and in-repo requirements. - Propose a minimal, correct story set that can be executed sequentially without merge pain. - Do NOT invent external requirements. DELIVERABLES - New markdown story files under stories_codex/_codex/ - Optional PROCESS.md under docs_codex/processes/_codex/ - Commit all created files with message: "add BMAD stories (canonical order)" NOW IMPLEMENT - Generate the story files. - Ensure numbering/order is correct and dependency-safe.