## Render when `LANG == "en"` - English guide ``` +----------------------------------+ | Multi-Agent Task Orchestrator | +----------------------------------+ 5 different input types supported - all enter the same flow: /multi-agent "{JIRA_KEY}-XXXXX" -> Jira ID /multi-agent "https://{JIRA_HOST}/browse/..." -> Jira URL /multi-agent "https://github.com/.../issues/N" -> GitHub Issue URL /multi-agent "#316" -> GitHub Issue # /multi-agent "bug/feature description" -> Free-text ------------------------------ How It Works (Phase 0 - Interactive Flow): 1. PARSE Classify input, fetch info from Jira/GitHub 2. PROJECT $HOME scanned, projects listed -> select by number 3. BRANCH Remote branches fetched (develop* first) -> select 4. DESIGN? Any Figma URL or screenshot? (optional) 5. BRANCH NAME bugfix/{jiraId} or feature/{jiraId} -> confirm 6. GIT ID Select identity from preferences (multiple supported) 7. INSTRUCTION Auto-detect if .instructions/ exists (figma etc.) 8. WORKSPACE Ask where the branch lives: worktree (default) or project root ------------------------------ Pipeline (after Phase 0) - shown as visual cards in terminal: Phase 0: Init -> The 8 steps above Phase 1: Plan -> Stack detection + codebase scan, task breakdown, architecture review, Plan Approval Gate (Opus; max 2 clarification rounds, interactive only - autopilot has nobody to ask) Phase 2: Dev -> TDD: test -> code -> build (Sonnet) + build queue, then the Verify exit gate: build, lint, tests, secrets. The build runs ONCE per run and its log carries into Review Phase 3: Review -> Parallel AI review + Fable triage, then the optional user test (Claude Code: Fable + Opus + Sonnet · Copilot: GPT-5.4 + Opus + Sonnet; the user test is the interactive half: branch, test in Xcode) Phase 4: Commit -> Commit -> push -> PR + issue body update (never auto-closes) Phase 5: Report -> Channels dispatcher (PR · Jira · Confluence · Wiki, multi-select) + internal capture (agent-log · knowledge · memory) Autopilot always pauses at the Phase 5 channels menu (30-min timeout → session ends cleanly). One pipeline. Every mode runs its whole phase set: Plan -> Dev(Sonnet) -> Review -> Commit -> Report What a run costs follows the evidence the task carries, not an answer taken before the evidence exists. Every step is logged. Error in any phase -> pause -> resume to continue. ------------------------------ Modes: Pipeline entries: /multi-agent "task" Worktree /multi-agent:autopilot "task" Worktree, no questions Phase 0 Step 5b asks where the branch lives; there is no flag for it. autopilot skips every confirmation INCLUDING the plan gate and the workspace question, and auto commit/PR - EXCEPT the Phase 5 channels menu, which always pauses. /multi-agent:resume [#id|jira-id] [autopilot] Pick up unfinished work: a stopped run, or a branch with no run behind it (Review → Commit/PR → Report) ------------------------------ Status & Resume: /multi-agent:status List tasks across all projects (id, phase, branch, state) /multi-agent:log [#id] Show task log (defaults to most recent) /multi-agent:resume [#id] Resume a paused or failed task from where it stopped /multi-agent:kill [#id] Stop task + delete worktree (logs preserved) /multi-agent:steer #id "..." Correct a running task without stopping it (applied at the next phase) /multi-agent:prune-logs Delete per-task project logs (audit + metrics kept; age/project/task filters, dry-run first) /multi-agent:prune-prompts Zero-base prompt review: measure always-on instruction footprint, propose keep/trial/delete per rule (report first, apply on approval) /multi-agent:garbage-collect Sweep run residue; --abandoned reaps stopped runs /multi-agent:purge Worktree + logs - full reset (double confirm) /multi-agent:autopilot-on Continuous mode on this machine: pick repos, labelled items run to a PR /multi-agent:autopilot-status Running, queued, awaiting an answer, today's PRs /multi-agent:autopilot-off Stop picking work up (running work finishes; --now stops it) /multi-agent:serve Contract server for a client: 127.0.0.1, token, stops when idle Post-Hoc & Side-Channel: /multi-agent:channels Post multi-channel report (Jira / Confluence / Wiki / PR description) /multi-agent:feedback "msg" Tell the maintainer something broke; only your text is sent /multi-agent:review Parallel review of a PR or branch diff; no URL -> pick open GitHub/Bitbucket PRs /multi-agent:review-jira Grade a Jira issue's pipeline-readiness -> comment the gaps on it /multi-agent:review-issue Grade a GitHub issue's pipeline-readiness -> comment the gaps on it /multi-agent:analysis ["name"] Feature-spec analysis (Figma + Swagger + Confluence + repos). Asks the standard first: global (23-section dev handoff) or corporate (IG→UC→FG requirements doc). Stack optional; References built from the evidence record /multi-agent:analysis-resolve [doc] Resolve Section 20 open questions of an analysis doc, one at a time with source-labeled candidates /multi-agent:research Research an item's gaps in tracker, docs and repo before asking; a gate decides what closed /multi-agent:review-analysis [doc] Review a written analysis; findings cite the rule they break /multi-agent:analysis-jira [doc] A final analysis -> a Jira story tree; coverage two-way, existing nodes skipped /multi-agent:complaint-analysis ["run-name"] Customer-complaint triage: Graylog evidence per trx/conv id + read-only repo correlation → root cause + fix plan, or core routing /multi-agent:build-optimize iOS-only Xcode build perf wrapper → benchmark + analyze + recommend-first .build-benchmark/optimization-plan.md /multi-agent:create-jira ["desc"] [figma-url] [swagger-url] Create a Jira Task/Bug/Story matching team conventions (type, sprint, preview + approval) /multi-agent:diff-explain Map a Phase 3 triage finding back to specific diff lines /multi-agent:estimate Effort range from similar closed work: P25-P75 per role, with the tickets behind it /multi-agent:scenario-audit Test scenarios vs the code: PRESENT / PARTIAL / ABSENT / UNCLEAR, each PRESENT on a resolving file:line /multi-agent:graph Build and query the repo code graph (symbols, imports, references), LLM-free /multi-agent:search Cross-task log search with smart ranking; --semantic queries triage corpus /multi-agent:scan Skill security scan against tiered pattern catalog /multi-agent:security-review Standalone defensive static security review: threat model, OWASP + CWE + CVSS findings, dep inventory. No live target. /multi-agent:doctor Would a run work here? Layout, prefs, credentials, hooks; exit code is the verdict /multi-agent:refactor Best practices + bug hunt + upstream drift + toolkit MCP research -> one plan, approval, dev + sync /multi-agent:refactor backlog Decide the friction already recorded about the pipeline itself, nothing re-derived /multi-agent:store-ready [repo] [flags] Pre-submission store readiness, iOS + Android, local-only; a skipped gate is never a pass; never ships a build. /multi-agent:testflight-validation [repo] [--ipa=|--archive=] iOS-pinned alias of :store-ready. Same three gates, one implementation. /multi-agent:ios-coding-standard [module] Audit an iOS module against the 99-rule coding-standard registry -> remediation plan + one-page onboarding summary -> hand off to /multi-agent. Read-only, never edits source. Setup & Maintenance: /multi-agent:setup First-run wizard - Keychain tokens + Git identity + language /multi-agent:language [en|tr] Show or set outputLanguage (promptLanguage stays English) /multi-agent:model [on|off] Turn the top model rung (fable) on or off; keeps costBudget.pricingModel in step /multi-agent:route-on Arm policy-driven model routing: strategy, scope, rules. Ships off /multi-agent:route-off Disarm routing and KEEP the rules, so route-on does not re-ask /multi-agent:route-status What is armed, which rule applies where, and the limit: subagents stay on Anthropic /multi-agent:stack [ids...] Enable stack plugin(s); multi-select: ids together (ios backend) or no arg -> native picker (common always on) /multi-agent:scaffold New project from the toolkit's scaffold skill; green before the first commit, no remote /multi-agent:sync Sync ecosystem (Claude Code + Copilot CLI + pipeline + website + toolkit MCP) /multi-agent:update Pull latest pipeline + reinstall + run migrations /multi-agent:uninstall Uninstall pipeline from every CLI (--all-data also clears settings, logs, memory + knowledge; tokens always intact) ------------------------------ Plugins & tools (called directly, no wrappers): Stack plugins ai--toolkit:, e.g. ai-ios-toolkit:create-component Catalog: ai--toolkit:help Enable per repo: /multi-agent:stack (common + analyst always on) Toolkit MCP multi-agent-toolkit - 80+ tools for a RUNNING app, no pipeline run needed: ios_* / android_* (launch, tap, screenshot, ui tree, logs, crashes, locale, dark mode), web_*, design_* (Design Check above), and the audits (ios_app_store_audit, ios_testflight_validate, android_apk_audit, *_accessibility_audit). Use them instead of guessing about on-screen state. Its registration survives uninstall. Not registered = silent no-op. ------------------------------ Routines (your own reusable jobs): /multi-agent:save [name] Save a recurring job as a reusable /multi-agent: command /multi-agent:routines List your saved routines and what each does /multi-agent:forget [name] Remove a saved routine ------------------------------ Interactive Launchers: /multi-agent:jira List MY open Jira issues → pick → branch → mode → go /multi-agent:issue List UNASSIGNED GitHub issues → pick → auto-assign → go Both follow the same flow after selection: 1. Branch selection (develop/release/main) 2. Autopilot: yes/no 3. Pipeline starts with the selected issue ------------------------------ UI Testing (standalone - not part of pipeline phases): /multi-agent:test Full simulator test (screenshot all screens) Fixed-scenario commands - no quoting, and they autocomplete off `test-`: /multi-agent:test-dark-mode Dark mode bug test /multi-agent:test-accessibility Accessibility audit (labels, tap targets, contrast) /multi-agent:test-dynamic-type Large text size test /multi-agent:test-screenshots [tr] App Store screenshot set in a locale (default tr) The scenario-tag form still works; each command above is its alias. /multi-agent:test "dark mode" | "accessibility" | "dynamic type" | "screenshot " `store-ready` is NOT a UI test: it validates a built package, on iOS and Android, through three gates per platform, and lives at /multi-agent:store-ready. The old /multi-agent:test "store-ready" [path] tag still works and hands off there. /multi-agent:bug-bash [scope] Parallel exploratory sessions by charter and persona, each with a budget; a finding is reported only when a repro test fails for its reason Uses xcrun simctl / adb (native, no external app needed). Booted simulator/emulator required. Auto-detects bundle ID from project. Manual Test (Phase 3 user-test step, standalone - Xcode hint flow): /multi-agent:manual-test Checkout task branch, print Xcode/SourceTree hints, /multi-agent:manual-test #N wait for your "ok" / "fix: ..." verdict. Design Check (mock-mode vs Figma, local-only): /multi-agent:design-check Whole-module audit: enumerate every state driver into a countable target set, build Debug in a worktree, drive each in mock mode, diff against its Figma frame, export HTML+PDF to ~/DesignChecks. /multi-agent:design-check Scope to one screen (e.g. order-status). /multi-agent:design-check Scope to one variant (e.g. order-status@expired). /multi-agent:design-check Scope to the frame(s) those URLs name. /multi-agent:design-check --resume Audit what the last run for this repo+module left unaudited. # COVERAGE GATE: every target is audited or skipped WITH a concrete reason; anything else reports # INCOMPLETE with the missing ids. "Needs a scenario/launch-arg" is not a reason, reaching it is the job. ------------------------------ Setup: /multi-agent:setup Interactive onboarding: 1. Dynamic Keychain Discovery - finds existing tokens by keyword 2. Token Mapping - maps found keys to services (preferences) 3. Jira Project Key - your team's key (MOBILE, APP, ENG, etc.) 4. Git Identity - corporate + personal (saved to preferences) 5. Missing Tokens - standard key names + ready-to-paste commands 6. Verify - re-scan, confirm ready ------------------------------ Key Features: Multi-Platform iOS, Android, Backend, Web - auto-detected Build Queue xcodebuild/Gradle serialized with lock file - parallel safe Stack Aware Auto-detects iOS/Android/Python/Node/Go/Docker, uses the enabled stack plugin's skills Det. Gates Build + lint + test + secret scan before AI review Multi-Repo Per-repo worktrees + identity, integration build before commit Identity Routing Git identity picked from the repo origin URL (corporate vs personal) Issue Safety Never auto-closes issues (4 approvals, GitHub + Jira) Store Compliance /multi-agent:test "store-ready" - 18-rule iOS audit (ITMS, Privacy Manifest, signing, debug leaks, IPv6, SDK) + 21-rule Android audit Bilingual EN + TR - outputLanguage toggles assistant explanations; promptLanguage is locked en Quality & Telemetry (advisory, on by default - flip prefs.global.* to disable): Diff Risk Score Phase 3 Step 1.75 ranks files before reviewer dispatch - heuristic, sub-second: security paths, migrations, no-test-change, complexity Test Gap Report Phase 3 surfaces public symbols added in this branch with no paired test Visual Evidence before/after stills + flow video on UI changes; Step 7.7 asks the depth Cost Breakdown Phase 5 appends per-phase tokens + estimated USD to agent-log.md Triage Memory Phase 5 ingests accepted/deferred/rejected findings into a repo corpus Prior-Art Lookup Phase 1 + Phase 3 query the corpus for similar findings, inject as context Per-Persona Dispatch reads `preferredModel` from the persona file; PHASE_MODEL_OVERRIDE overrides per call; ladder fable -> opus -> sonnet -> haiku ------------------------------ Examples: # Full pipeline with Jira (key configured during setup) /multi-agent "MOBILE-12345" # Interactive launchers - browse and pick /multi-agent:jira # Pick from my Jira issues /multi-agent:issue # Pick from unassigned GitHub issues # Free-text - any platform /multi-agent:autopilot "LoginView dark mode fix" # iOS /multi-agent "Add pagination to /api/users endpoint" # Backend /multi-agent "Fix bottom nav recomposition in HomeScreen" # Android /multi-agent "Responsive layout broken on tablet" # Web # Review current diff only /multi-agent:review # Create a Jira issue that matches team conventions (preview + approval before create) /multi-agent:create-jira "Profile screen empty state" https://figma.com/design/abc?node-id=1-2 ------------------------------ Logs: $HOME/.claude/logs/multi-agent/{project}/{task-id}/ ```