[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7C1Q13SVM980TYZ4SE34FPC":3,"skill-same:01M7C1Q13SVM980TYZ4SE34FPC":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7C1Q13SVM980TYZ4SE34FPC","iterate",null,"Use when implementing a project with a large, comprehensive, or ambiguous spec — extracts requirements with proof obligations, defines a walking skeleton with its first journey scenario, then loops through audited sprints that continuously build a behavior evidence corpus. Completion means passing evidence, not just finished stories.","zak-keown\u002Fmoe","zak-keown","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F320717241?v=4","01M7BZFXME0NW6TQX0H7AZM2KS","01M7BZFXN32VR0JN5AATRD53H4",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002F.pi\u002Fskills\u002Fiterate\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: iterate\ndescription: Use when implementing a project with a large, comprehensive, or ambiguous spec — extracts requirements with proof obligations, defines a walking skeleton with its first journey scenario, then loops through audited sprints that continuously build a behavior evidence corpus. Completion means passing evidence, not just finished stories.\ntriggers: >-\n  Load when the spec is large (10+ files, 100+ requirements), ambiguous,\n  and you need audited sprints with behavior evidence. Do NOT load for:\n  well-scoped features with clear requirements (`write-plan` then\n  `sdd`), single tasks, bug fixes, or projects\n  where upfront planning has not yet been attempted.\n---\n\n# Iterative Development\n\n## Overview\n\nOrchestrator for the iterate plugin. Drives the full autonomous lifecycle: extract requirements with proof obligations and behavior scenarios from human spec collateral, define a walking skeleton that passes its first journey scenario, then loop through audited sprints that continuously build a reusable behavior evidence corpus. Completion means the product has passing behavior evidence at the correct seam for every externally observable requirement — not just that stories are marked done. Every evaluative gate uses parallel adversarial review (PAR).\n\nThis is an alternative to the `write-plan` → `sdd` flow for projects where the upfront-planning approach would lose the plot.\n\n## When to Use\n\n- Spec is large, comprehensive, or ambiguous (10+ files, 100+ requirements)\n- You need the product to be in a working, testable state at every iteration boundary\n- You want an autonomous audited loop rather than a single upfront plan\n- The write-plan flow has lost the plot on this project before\n\nDo NOT use for a `patch` or a `change` (see `brainstorming`) — the `write-plan` → `sdd` flow is simpler and more appropriate for those depths.\n\n## The Autonomous Loop\n\n### Bootstrap (first invocation)\n\n1. Check `docs\u002Fmoe\u002Fiterations\u002F` for existing state. If no state exists, call\n   `moe jig iterations init` to scaffold the directory, then proceed with\n   requirements extraction. If `moe-jig` is not on PATH, create the\n   directories and files manually. Existing state selects an explicit\n   continuation; no state selects new work.\n2. Establish one project workspace with `use-worktrees`. Record its\n   `WORKSPACE_ID`, `BASE_BRANCH`, `BASE_SHA`, and work branch in\n   `docs\u002Fmoe\u002Fiterations\u002Fworkspace.md`, and keep all ITER-NNNN iterations on\n   that project branch. For continuation, verify that record matches this\n   checkout and then skip to **Resume** below. Iterations do not create sibling\n   project branches; task workers may use temporary worktrees that\n   `implement-tasks` integrates back into this workspace.\n3. Invoke `extract-requirements` on the human-provided spec path.\n   - Chunks the spec, classifies by taxonomy (journeys → E2E, domains → integration, etc.)\n   - Dispatches parallel extraction subagents that produce stories with proof obligations AND behavior scenarios\n   - Aggregates stories into per-epic files, scenarios into behavior-scenarios.md\n   - Builds coverage ledger with both story AND scenario coverage\n   - Produces `docs\u002Fmoe\u002Fiterations\u002Frequirements\u002F`, `docs\u002Fmoe\u002Fiterations\u002Fbehavior-scenarios.md`, `docs\u002Fmoe\u002Fiterations\u002Fbehavior-corpus.md`\n4. Invoke `scope-core` on the resulting backlog.\n   - Defines the walking skeleton iteration (ITER-0000) + ordered follow-on iterations\n   - Runs citation check + PAR scope review\n   - Produces `docs\u002Fmoe\u002Fiterations\u002Froadmap.md`\n   - Walking skeleton must close at least one journey scenario (not just compile)\n   - Applies story splitting when stories have heterogeneous-dependency ACs\n\n### Main loop\n\n```\nwhile True:\n    check_for_human_interrupt()\n\n    if not roadmap has pending iterations:\n        if last audit was clean:\n            run final behavior-evidence audit (see below)\n            if behavior audit clean:\n                break  # done\n            # else: audit found uncovered surfaces or weak evidence, new iterations added\n        # else: audit found gaps, new iterations were added, continue\n\n    run next iteration:\n        - run-iteration (sentinel baseline → scope review → decompose code + evidence tasks → implement-tasks → impacted + sentinel scenario runs → wrap up)\n\n    audit:\n        - audit-progress (PAR paired auditors, three-tier: deep evidence + impacted behavior + sentinel corpus)\n        - if gaps: append to backlog, revise roadmap, continue\n        - if clean: mark last_audit_clean, continue\n```\n\n### Final behavior-evidence audit\n\nBefore declaring the project complete, verify that the product has adequate behavior evidence — not just that all stories are marked done:\n\n1. List every major user-facing surface from the original spec (settings panes, UI flows, CLI commands, journeys, etc.)\n2. For each surface, verify that:\n   - Corresponding stories exist AND are implemented\n   - Corresponding scenarios exist AND have passing evidence at the correct seam\n   - Journey scenarios that cross multiple surfaces are passing E2E\n3. Check the behavior corpus index for completeness:\n   - Every journey spec file has at least one JOURNEY-NNNN scenario\n   - Every scenario has a non-TBD execution command\n   - All sentinel scenarios pass\n4. Flag any surface with:\n   - No corresponding story (extraction under-scoped)\n   - No corresponding scenario (evidence gap)\n   - Evidence at a weaker seam than the requirement demands\n   - Manual-residual scenarios that could be automated\n5. If gaps found: create new stories\u002Fscenarios\u002Fiterations, continue the loop\n\nThe final question is: \"Can the system point to passing behavior evidence for every externally observable requirement the spec describes?\" Not: \"Are the stories done?\"\n\nWhen the answer is yes, invoke `finish-branch` exactly once,\nafter the final behavior-evidence audit. This is the project-level integration\nboundary: verify the full suite, present the normal merge\u002FMR\u002Fkeep decision, and\nperform the selected cleanup. Do not finish or rebranch between iterations.\n\n### Resume (re-invocation with existing state)\n\nAll process state lives in artifact files:\n- `docs\u002Fmoe\u002Fiterations\u002Frequirements\u002F` (backlog with story status and proof obligations)\n- `docs\u002Fmoe\u002Fiterations\u002Fbehavior-scenarios.md` (scenario cards with stable IDs)\n- `docs\u002Fmoe\u002Fiterations\u002Fbehavior-corpus.md` (execution index)\n- `docs\u002Fmoe\u002Fiterations\u002Froadmap.md` (iteration plan with status)\n- `docs\u002Fmoe\u002Fiterations\u002Fiteration-log.md` (completed iteration history)\n- `docs\u002Fmoe\u002Fiterations\u002Fworkspace.md` (project workspace identity, base branch\u002FSHA, and work branch)\n\nOn re-invocation: read `roadmap.md`, verify the recorded project workspace\nidentity and base, find the next pending iteration, and continue from there.\nThere is no ephemeral in-memory state to recover. The command \"continue\niterative development with the existing plan\" always works from the matching\nproject workspace; an unrelated linked worktree must not adopt it implicitly.\n\nIf the orchestrator crashed mid-iteration, the partially-completed iteration's git commits are preserved. On resume, the next un-started iteration picks up. If the in-progress iteration left the code in a broken state, treat it as a gap — the audit will catch it and add corrective work.\n\n## Human Interrupt Protocol\n\nThe loop runs without human intervention. The only way the human injects new information mid-run is by interrupting between iterations.\n\n**How it works:**\n- The human types the update into the chat session (\"we dropped feature X\", \"the spec changed, re-read specs\u002Ffoo.md\", \"add a new requirement for Y\")\n- The orchestrator notices the interrupt at the **next iteration boundary** — after the current iteration's audit completes, before the next iteration starts\n- At the boundary: invoke `extract-requirements` in incremental mode on the changed spec files, merge new\u002Frevised story cards into the backlog, revise the roadmap if changes invalidate downstream iterations, then resume\n\n**Guarantees:**\n- Changes during mid-iteration do NOT disrupt in-progress work. The current iteration completes first.\n- The orchestrator never silently drops an interrupt. If ambiguous, ask for clarification before resuming.\n- Existing story IDs are preserved across re-extraction. Removed stories flip to `deferred`, not deleted.\n\n**What does NOT trigger interrupt processing:**\n- The orchestrator does not poll the filesystem for spec changes\n- The orchestrator does not ask \"anything to change?\" between iterations\n- Human presence is not required at iteration boundaries\n\n## Progress Reporting\n\nThe autonomous loop may run for hours. Two progress mechanisms ensure visibility without requiring interruption:\n\n**1. Progress file:** Write `docs\u002Fmoe\u002Fiterations\u002Fprogress.md` at each phase transition:\n\n```markdown\n# Progress\n\n**Phase:** implementing ITER-0003\n**Task:** 4\u002F7 (CleanupPipeline integration)\n**Iterations:** 3\u002F18 done, 15 pending\n**Sentinel corpus:** 10\u002F10 passing\n**Last event:** 2026-04-11T14:23:00Z — Task 3 committed\n```\n\nUpdate this file at: iteration start, each task completion, iteration wrap-up, audit start\u002Fend. Overwrite (not append) — it's a snapshot of current state, not a log. Call `moe jig progress update --phase \u003Cphase> --task \u003Ctask>` at each phase transition. The command overwrites `progress.md` with the correct format. If `moe-jig` is not on PATH, write the file manually using the exact format above.\n\n**2. Git log:** Every task produces a commit. The commit history is a detailed progress trail. A human can check `git log --oneline` for fine-grained status without interrupting the loop.\n\n## Skill Precedence\n\nWhen running autonomously, this orchestrator takes precedence over interactive-gate skills (e.g., `brainstorming` which requires design approval before implementation). The iterate process has its own design gates (scope review, PAR) that replace interactive approval. Do not block on skills that assume a human is present to approve each step.\n\n## Escalation Policy\n\n**Catastrophe-only.** The loop is autonomous. Human escalation is reserved for total failure — the plugin cannot make any forward progress at all.\n\nThese do NOT trigger escalation:\n- A reviewer finding issues (those become fix work)\n- An audit finding gaps (those become new iterations)\n- An implementer reporting BLOCKED on a task (try: more context, more capable model, smaller task)\n- Ambiguity in the spec (make a reasonable judgment call, document it in the iteration log)\n- Difficulty or slow progress (keep going)\n\nThe orchestrator does NOT prompt \"should I continue?\" between iterations.\n\n## Skill Invocation Reference\n\n| Phase | Skill | What it does |\n|---|---|---|\n| Extract | `extract-requirements` | Chunk → parallel extract → aggregate → `requirements\u002F` |\n| Scope | `scope-core` | Walking skeleton + iterations → `roadmap.md` (with PAR scope review) |\n| Implement | `run-iteration` | Scope review → decompose → `implement-tasks` → wrap up |\n| Task execution | `implement-tasks` | Per-task: implementer → PAR spec review → PAR quality review |\n| Audit | `audit-progress` | PAR paired auditors, two-tier (deep + sweep) |\n\n## Artifact Location\n\nAll plugin artifacts live in `docs\u002Fmoe\u002Fiterations\u002F`. Never modify the human's spec collateral.\n\n| File | Purpose |\n|---|---|\n| `requirements\u002F` | Backlog: story cards + epics with stable IDs and proof obligations |\n| `behavior-scenarios.md` | Behavior contracts: reusable scenario cards with stable IDs |\n| `behavior-corpus.md` | Execution index: scenario → seam → cadence → command |\n| `roadmap.md` | Sprint plan: ordered iterations with impacted scenarios |\n| `iteration-log.md` | Sprint history: what each iteration delivered + scenarios added |\n| `progress.md` | Live snapshot: current phase, task, iteration counts, sentinel status |\n| `workspace.md` | Project workspace identity, immutable fork SHA, intended merge branch, and work branch |\n\n## Quality Gates\n\nEvery evaluative gate uses parallel adversarial review (PAR):\n- Pre-iteration scope review (citation + scope-creep + boxing-in + scenario coverage + story splitting)\n- Pre-iteration sentinel corpus baseline\n- Per-task spec-compliance review with evidence quality check\n- Per-task code-quality review with boxing-in + corpus contribution check\n- Post-iteration impacted + sentinel scenario runs\n- Per-sprint audit (deep evidence + impacted behavior + sentinel corpus)\n\nResolve [skills\u002F_shared\u002Fparallel-adversarial-review.md](..\u002F_shared\u002Fparallel-adversarial-review.md) relative to this loaded document for the PAR methodology.\n","# Iterative Development\n\n## Overview\n\nOrchestrator for the iterate plugin. Drives the full autonomous lifecycle: extract requirements with proof obligations and behavior scenarios from human spec collateral, define a walking skeleton that passes its first journey scenario, then loop through audited sprints that continuously build a reusable behavior evidence corpus. Completion means the product has passing behavior evidence at the correct seam for every externally observable requirement — not just that stories are marked done. Every evaluative gate uses parallel adversarial review (PAR).\n\nThis is an alternative to the `write-plan` → `sdd` flow for projects where the upfront-planning approach would lose the plot.\n\n## When to Use\n\n- Spec is large, comprehensive, or ambiguous (10+ files, 100+ requirements)\n- You need the product to be in a working, testable state at every iteration boundary\n- You want an autonomous audited loop rather than a single upfront plan\n- The write-plan flow has lost the plot on this project before\n\nDo NOT use for a `patch` or a `change` (see `brainstorming`) — the `write-plan` → `sdd` flow is simpler and more appropriate for those depths.\n\n## The Autonomous Loop\n\n### Bootstrap (first invocation)\n\n1. Check `docs\u002Fmoe\u002Fiterations\u002F` for existing state. If no state exists, call\n   `moe jig iterations init` to scaffold the directory, then proceed with\n   requirements extraction. If `moe-jig` is not on PATH, create the\n   directories and files manually. Existing state selects an explicit\n   continuation; no state selects new work.\n2. Establish one project workspace with `use-worktrees`. Record its\n   `WORKSPACE_ID`, `BASE_BRANCH`, `BASE_SHA`, and work branch in\n   `docs\u002Fmoe\u002Fiterations\u002Fworkspace.md`, and keep all ITER-NNNN iterations on\n   that project branch. For continuation, verify that record matches this\n   checkout and then skip to **Resume** below. Iterations do not create sibling\n   project branches; task workers may use temporary worktrees that\n   `implement-tasks` integrates back into this workspace.\n3. Invoke `extract-requirements` on the human-provided spec path.\n   - Chunks the spec, classifies by taxonomy (journeys → E2E, domains → integration, etc.)\n   - Dispatches parallel extraction subagents that produce stories with proof obligations AND behavior scenarios\n   - Aggregates stories into per-epic files, scenarios into behavior-scenarios.md\n   - Builds coverage ledger with both story AND scenario coverage\n   - Produces `docs\u002Fmoe\u002Fiterations\u002Frequirements\u002F`, `docs\u002Fmoe\u002Fiterations\u002Fbehavior-scenarios.md`, `docs\u002Fmoe\u002Fiterations\u002Fbehavior-corpus.md`\n4. Invoke `scope-core` on the resulting backlog.\n   - Defines the walking skeleton iteration (ITER-0000) + ordered follow-on iterations\n   - Runs citation check + PAR scope review\n   - Produces `docs\u002Fmoe\u002Fiterations\u002Froadmap.md`\n   - Walking skeleton must close at least one journey scenario (not just compile)\n   - Applies story splitting when stories have heterogeneous-dependency ACs\n\n### Main loop\n\n```\nwhile True:\n    check_for_human_interrupt()\n\n    if not roadmap has pending iterations:\n        if last audit was clean:\n            run final behavior-evidence audit (see below)\n            if behavior audit clean:\n                break  # done\n            # else: audit found uncovered surfaces or weak evidence, new iterations added\n        # else: audit found gaps, new iterations were added, continue\n\n    run next iteration:\n        - run-iteration (sentinel baseline → scope review → decompose code + evidence tasks → implement-tasks → impacted + sentinel scenario runs → wrap up)\n\n    audit:\n        - audit-progress (PAR paired auditors, three-tier: deep evidence + impacted behavior + sentinel corpus)\n        - if gaps: append to backlog, revise roadmap, continue\n        - if clean: mark last_audit_clean, continue\n```\n\n### Final behavior-evidence audit\n\nBefore declaring the project complete, verify that the product has adequate behavior evidence — not just that all stories are marked done:\n\n1. List every major user-facing surface from the original spec (settings panes, UI flows, CLI commands, journeys, etc.)\n2. For each surface, verify that:\n   - Corresponding stories exist AND are implemented\n   - Corresponding scenarios exist AND have passing evidence at the correct seam\n   - Journey scenarios that cross multiple surfaces are passing E2E\n3. Check the behavior corpus index for completeness:\n   - Every journey spec file has at least one JOURNEY-NNNN scenario\n   - Every scenario has a non-TBD execution command\n   - All sentinel scenarios pass\n4. Flag any surface with:\n   - No corresponding story (extraction under-scoped)\n   - No corresponding scenario (evidence gap)\n   - Evidence at a weaker seam than the requirement demands\n   - Manual-residual scenarios that could be automated\n5. If gaps found: create new stories\u002Fscenarios\u002Fiterations, continue the loop\n\nThe final question is: \"Can the system point to passing behavior evidence for every externally observable requirement the spec describes?\" Not: \"Are the stories done?\"\n\nWhen the answer is yes, invoke `finish-branch` exactly once,\nafter the final behavior-evidence audit. This is the project-level integration\nboundary: verify the full suite, present the normal merge\u002FMR\u002Fkeep decision, and\nperform the selected cleanup. Do not finish or rebranch between iterations.\n\n### Resume (re-invocation with existing state)\n\nAll process state lives in artifact files:\n- `docs\u002Fmoe\u002Fiterations\u002Frequirements\u002F` (backlog with story status and proof obligations)\n- `docs\u002Fmoe\u002Fiterations\u002Fbehavior-scenarios.md` (scenario cards with stable IDs)\n- `docs\u002Fmoe\u002Fiterations\u002Fbehavior-corpus.md` (execution index)\n- `docs\u002Fmoe\u002Fiterations\u002Froadmap.md` (iteration plan with status)\n- `docs\u002Fmoe\u002Fiterations\u002Fiteration-log.md` (completed iteration history)\n- `docs\u002Fmoe\u002Fiterations\u002Fworkspace.md` (project workspace identity, base branch\u002FSHA, and work branch)\n\nOn re-invocation: read `roadmap.md`, verify the recorded project workspace\nidentity and base, find the next pending iteration, and continue from there.\nThere is no ephemeral in-memory state to recover. The command \"continue\niterative development with the existing plan\" always works from the matching\nproject workspace; an unrelated linked worktree must not adopt it implicitly.\n\nIf the orchestrator crashed mid-iteration, the partially-completed iteration's git commits are preserved. On resume, the next un-started iteration picks up. If the in-progress iteration left the code in a broken state, treat it as a gap — the audit will catch it and add corrective work.\n\n## Human Interrupt Protocol\n\nThe loop runs without human intervention. The only way the human injects new information mid-run is by interrupting between iterations.\n\n**How it works:**\n- The human types the update into the chat session (\"we dropped feature X\", \"the spec changed, re-read specs\u002Ffoo.md\", \"add a new requirement for Y\")\n- The orchestrator notices the interrupt at the **next iteration boundary** — after the current iteration's audit completes, before the next iteration starts\n- At the boundary: invoke `extract-requirements` in incremental mode on the changed spec files, merge new\u002Frevised story cards into the backlog, revise the roadmap if changes invalidate downstream iterations, then resume\n\n**Guarantees:**\n- Changes during mid-iteration do NOT disrupt in-progress work. The current iteration completes first.\n- The orchestrator never silently drops an interrupt. If ambiguous, ask for clarification before resuming.\n- Existing story IDs are preserved across re-extraction. Removed stories flip to `deferred`, not deleted.\n\n**What does NOT trigger interrupt processing:**\n- The orchestrator does not poll the filesystem for spec changes\n- The orchestrator does not ask \"anything to change?\" between iterations\n- Human presence is not required at iteration boundaries\n\n## Progress Reporting\n\nThe autonomous loop may run for hours. Two progress mechanisms ensure visibility without requiring interruption:\n\n**1. Progress file:** Write `docs\u002Fmoe\u002Fiterations\u002Fprogress.md` at each phase transition:\n\n```markdown\n# Progress\n\n**Phase:** implementing ITER-0003\n**Task:** 4\u002F7 (CleanupPipeline integration)\n**Iterations:** 3\u002F18 done, 15 pending\n**Sentinel corpus:** 10\u002F10 passing\n**Last event:** 2026-04-11T14:23:00Z — Task 3 committed\n```\n\nUpdate this file at: iteration start, each task completion, iteration wrap-up, audit start\u002Fend. Overwrite (not append) — it's a snapshot of current state, not a log. Call `moe jig progress update --phase \u003Cphase> --task \u003Ctask>` at each phase transition. The command overwrites `progress.md` with the correct format. If `moe-jig` is not on PATH, write the file manually using the exact format above.\n\n**2. Git log:** Every task produces a commit. The commit history is a detailed progress trail. A human can check `git log --oneline` for fine-grained status without interrupting the loop.\n\n## Skill Precedence\n\nWhen running autonomously, this orchestrator takes precedence over interactive-gate skills (e.g., `brainstorming` which requires design approval before implementation). The iterate process has its own design gates (scope review, PAR) that replace interactive approval. Do not block on skills that assume a human is present to approve each step.\n\n## Escalation Policy\n\n**Catastrophe-only.** The loop is autonomous. Human escalation is reserved for total failure — the plugin cannot make any forward progress at all.\n\nThese do NOT trigger escalation:\n- A reviewer finding issues (those become fix work)\n- An audit finding gaps (those become new iterations)\n- An implementer reporting BLOCKED on a task (try: more context, more capable model, smaller task)\n- Ambiguity in the spec (make a reasonable judgment call, document it in the iteration log)\n- Difficulty or slow progress (keep going)\n\nThe orchestrator does NOT prompt \"should I continue?\" between iterations.\n\n## Skill Invocation Reference\n\n| Phase | Skill | What it does |\n|---|---|---|\n| Extract | `extract-requirements` | Chunk → parallel extract → aggregate → `requirements\u002F` |\n| Scope | `scope-core` | Walking skeleton + iterations → `roadmap.md` (with PAR scope review) |\n| Implement | `run-iteration` | Scope review → decompose → `implement-tasks` → wrap up |\n| Task execution | `implement-tasks` | Per-task: implementer → PAR spec review → PAR quality review |\n| Audit | `audit-progress` | PAR paired auditors, two-tier (deep + sweep) |\n\n## Artifact Location\n\nAll plugin artifacts live in `docs\u002Fmoe\u002Fiterations\u002F`. Never modify the human's spec collateral.\n\n| File | Purpose |\n|---|---|\n| `requirements\u002F` | Backlog: story cards + epics with stable IDs and proof obligations |\n| `behavior-scenarios.md` | Behavior contracts: reusable scenario cards with stable IDs |\n| `behavior-corpus.md` | Execution index: scenario → seam → cadence → command |\n| `roadmap.md` | Sprint plan: ordered iterations with impacted scenarios |\n| `iteration-log.md` | Sprint history: what each iteration delivered + scenarios added |\n| `progress.md` | Live snapshot: current phase, task, iteration counts, sentinel status |\n| `workspace.md` | Project workspace identity, immutable fork SHA, intended merge branch, and work branch |\n\n## Quality Gates\n\nEvery evaluative gate uses parallel adversarial review (PAR):\n- Pre-iteration scope review (citation + scope-creep + boxing-in + scenario coverage + story splitting)\n- Pre-iteration sentinel corpus baseline\n- Per-task spec-compliance review with evidence quality check\n- Per-task code-quality review with boxing-in + corpus contribution check\n- Post-iteration impacted + sentinel scenario runs\n- Per-sprint audit (deep evidence + impacted behavior + sentinel corpus)\n\nResolve [skills\u002F_shared\u002Fparallel-adversarial-review.md](..\u002F_shared\u002Fparallel-adversarial-review.md) relative to this loaded document for the PAR methodology.\n",{"name":5,"triggers":23,"description":7},"Load when the spec is large (10+ files, 100+ requirements), ambiguous, and you need audited sprints with behavior evidence. Do NOT load for: well-scoped features with clear requirements (`write-plan` then `sdd`), single tasks, bug fixes, or projects where upfront planning has not yet been attempted.","main","728894f3f85fbd4e788f015ddfd72bb6155d5dd5","plugins\u002Fmoe\u002F.pi\u002Fskills\u002Fiterate",1,"https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe","https:\u002F\u002Fgithub.com\u002Fzak-keown","2026-09-05 06:29:06",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add zak-keown\u002Fmoe --skill iterate",{"tool":37,"command":38},"bunx","bunx skills add zak-keown\u002Fmoe --skill iterate",{"tool":40,"command":41},"pnpm","pnpm dlx skills add zak-keown\u002Fmoe --skill iterate","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002F.pi\u002Fskills\u002Fiterate\u002FSKILL.md",true,[],[46,53,60,67,74,81],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":18},"01M7C1Z5362Y8NVRZ69HWP0ZD0","write-skill","Use when creating new skills, editing existing skills, or verifying skills work before deployment","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-skill\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":18},"01M7C1YZ9966WK9R6ZRPCGWF3G","write-plan","Use when you have a spec or requirements for a multi-step task, before touching code","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-plan\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":18},"01M7C1YMR4SJD7BPBBPXZCVGVH","write-clearly","Apply Strunk's timeless writing rules to ANY prose humans will read—documentation, commit messages, error messages, explanations, reports, or UI text. Makes your writing clearer, stronger, and more professional.","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwrite-clearly\u002FSKILL.md",[],[],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":18},"01M7C1YH9805YSM0N20VQAVSXF","windows-vm","Create, manage, or connect to a headless Windows 11 VM running in Docker with SSH access. Use when the user wants to spin up, stop, restart, or SSH into a Windows VM.","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fwindows-vm\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":18},"01M7C1YD87SCNEJ2C9B8HFR2NQ","verify-completion","Use when about to claim work is complete, fixed, or passing, before committing or creating MRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fverify-completion\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":84,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":18},"01M7C1YARDEHT0TJ9BRG30ZCF4","using-moe","Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions","https:\u002F\u002Fgithub.com\u002Fzak-keown\u002Fmoe\u002Fblob\u002Fmain\u002Fplugins\u002Fmoe\u002Fskills\u002Fusing-moe\u002FSKILL.md",[],[]]