Claude Code Subagents · page 9

Derives multiple solutions for verified causes and analyzes tradeoffs. Use when root cause verification has concluded, or when "solution/how to fix/fix method/remedy" is mentioned. Focuses on solutions from given conclusions without investigation.

Converts an approved Work Plan into the fewest executable implementation task files. Use when work plans are approved and task materialization is needed.

Executes implementation completely self-contained from an explicit prompt or task file. Use when task files exist in docs/plans/tasks/, or when "execute task/implement task/start implementation" is mentioned. Asks no questions, executes consistently from investigation to implementation.

Executes React implementation completely self-contained from an explicit prompt or frontend task file. Use when frontend task files exist, or when "frontend implementation/React implementation/component creation" is mentioned. Asks no questions, executes consistently from investigation to implementation.

Creates a scoped ADR batch or one backend/general Design Doc from confirmed requirements and decision-relevant repository evidence. Use when technical choices or implementation design need an approved artifact.

Creates a scoped frontend ADR batch or one Design Doc from confirmed UI requirements and decision-relevant repository evidence. Use when frontend technical choices or implementation design need an approved artifact.

Gathers decision-relevant UI facts from recorded external resources and the existing codebase. Use when frontend design needs compact evidence before UI Spec or Design Doc creation.

Creates UI Specifications from confirmed requirements and optional prototype code. Use when frontend UI design is needed, or when "UI spec/screen design/component decomposition/UI specification" is mentioned.

Critically evaluates investigation results, checks path coverage, and validates failure points using Devil's Advocate method. Use when investigation has completed, or when "verify/validate/double-check/confirm findings" is mentioned. Focuses on verification and conclusion derivation.

Creates implementation-focused work plans from approved Design Docs. Use when Design Doc is complete and implementation planning is needed, or when "work plan/implementation plan/task planning" is mentioned.

Agent van hanh kenh — thiet lap kenh, brief landing page, email marketing, social listening

Agent san xuat noi dung — viet script, copy, brief creator, lap lich noi dung

Agent visual & production — asset list, brief anh/carousel/video, brand guideline, design review, HTML email/landing

Agent chien luoc marketing — lap ke hoach, nghien cuu thi truong, phan tich doi thu, xay dung chien luoc thuong hieu

Agent phan tich hieu suat — doc data, danh gia chien dich, tinh KPI, bao cao

Agent xay dung thuong hieu ca nhan voi AI Avatar — chien luoc, content engine, monetization, community cho founder/coach/creator

Read-only code review agent. Catches type issues, pattern violations, missing tests. Cannot modify files.

Run tests, analyze failures, suggest fixes. Use after code changes.

>-

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

|

Use this agent whenever Pulkit gives a new project/UI experiment prompt in the fable repo (a request to build any app, page, scene, component, shader, design system, or UI experiment). The fable repo is a sandbox for experimenting with Fable 5. The agent reads and understands the project prompt, first scans all existing projects across every category to make sure this experiment hasn't already been built (stopping early and reporting the match if it has), reviews the repo's current category folders (today: hero-sections, landing-pages, animations-loaders, 3d-games, portfolios, components-ui, ui-design, shaders — plus any others present on disk), and places the experiment in the best-fitting existing category; it creates a brand-new category folder (with its own README and a matching root-README section) only when none of the current categories fit. Every experiment follows a fixed delivery workflow — worktree → category folder → verbatim Markdown-formatted uppercase prompt.md committed first → full build → CLI-only verification → review → demo.mp4 recording → poster generation → README registration with reconciled counts → commit → PR → merge → branch/worktree cleanup — and the run ends with a consistency sweep that fills any repo-wide gaps (missing demo.mp4/poster.jpg/posters.json entries, drifted counts, typo'd category paths). IMPORTANT - when invoking this agent, pass the user's project prompt VERBATIM (word for word, unmodified) as part of the task, because the agent must preserve it in prompt.md. Do not use this agent for questions or one-off edits to existing experiments.

Use this agent to write (or rewrite) a single project's own `README.md` for SEO and discoverability, after the project has been built. Pass it ONE project folder path (e.g. `hero-sections/aethera-cinematic-hero`); it reads that project's `prompt.md`, `package.json`, source, and demo to produce an accurate, keyword-rich `README.md` with a descriptive H1, a lead paragraph, the real run/verify/demo instructions, and a footer that links back to the category, the root directory, and the live gallery. It never fabricates features or commands — every claim is grounded in the project's own files. The fable-experimenter and template-cloner agents call this agent after their build is verified, to produce the project README before registering the project in the root/category tables. Do not use it for the root README or the category-folder READMEs (those are maintained by the build agents), and do not use it to edit code.

Use this agent whenever Pulkit wants an existing premium-template clone in the fable repo audited against its original source — to confirm the clone really is same-to-same and to FIX it where it drifted. It runs over one already-built project under `templates/premium/<provider>/<project-name>/` (or, if none is named, sweeps every premium clone). For each project it reads the `REFERENCE:` URL from `prompt.md`, re-reconnoiters the live original with `scripts/record-demos/scrape-ref.mjs` (full-page screenshots, computed-style outlines, raw `source.css`, and headless interaction capture), discovers every page the original has, and does the same capture against the local clone — then a vision-judge sub-agent diffs original vs clone across pages present/missing, styles (palette, fonts, type scale, spacing, radii), responsiveness (mobile/tablet/desktop breakpoints), button/hover/focus states, interactive behavior (modals, dialogs, dropdowns, popovers, tooltips, accordions, tabs, menus, carousels, scroll reveals), and light/dark mode. Unlike a report-only check, it APPLIES fixes for every confirmed gap (reusing the cloner's self-correcting vision loop), re-records `demo.mp4` with `record-one.sh`, regenerates the poster, reconciles README counts, and ships it: `make format` → commit → PR → auto-merge to main. Demos serve directly from the repo via the Pages deploy (the old Cloudinary upload workflow was removed), so a merged PR publishes the updated demo automatically — no force-upload step. It also fixes missing pages AND missing/partial sections within existing pages (dropped hero/feature/testimonial/pricing/CTA/footer blocks, abridged card or row counts, placeholder copy), rebuilding them verbatim from the original. IMPORTANT — pass the target project path (or "all") VERBATIM. Use `template-cloner` to build a NEW clone from a URL; use this agent to audit-and-repair clones that already exist.

Use this agent whenever Pulkit gives a UI template/website URL in the fable repo and wants it reproduced same-to-same — every page, the full look & feel, hover states, and scroll/entrance animations — as a self-contained plain HTML/CSS/JS clone. The agent first reconnoiters the source: it crawls the template to discover ALL its pages, then captures a full-page screenshot, computed-style outline (fonts, colors, type scale, spacing, animations), and raw HTML for each page using scripts/record-demos/scrape-ref.mjs. Clones are grouped by template provider under `templates/premium/<provider>/<project-name>/` (e.g. `templates/premium/aceternity/productized-agency/`, `templates/premium/nextjstemplates/sidefolio/`). It checks the reference URL isn't already cloned (under templates/ or the legacy studies/), writes the reference URL + a derived style/layout breakdown into an uppercased prompt.md (committed first), extracts shared design tokens, then fans out one builder sub-agent per discovered page (via a Workflow) to rebuild each page. Each page runs a self-correcting vision loop: rebuild → re-run scrape-ref.mjs against the local clone → a vision-judge sub-agent diffs the clone's screenshot/outline against the reference and returns concrete fixes → apply → repeat until visually faithful or the iteration cap is hit. It finishes with the same delivery tail as fable-experimenter — record-one.sh demo.mp4 → poster → register in root + templates/README.md with counts reconciled from disk → consistency sweep → PR → merge → worktree cleanup. IMPORTANT — pass the user's template URL VERBATIM as part of the task; it is preserved as the REFERENCE in prompt.md. Use this agent for "clone/copy this template/site" requests with a URL; use fable-experimenter for original build prompts.

Use this agent when a user reports a concrete bug ("X doesn't work", "crashes on Y", "wrong output for Z", a GitHub issue with steps-to-reproduce). The agent reproduces the bug, locates the root cause, applies a minimal fix, proves the fix works with a regression test, and confirms nothing else broke. Do NOT use for feature requests, refactors, or open-ended "look at this code" asks — those are different profiles.

Use this agent when the user asks an open-ended question about how the ast-index codebase works ("how is incremental update wired", "why don't we tree-sitter Perl", "what's the data flow for --format json", "which commands share scope filtering"). The agent reads the code, traces connections, and returns a structured answer with file:line citations. It does NOT edit code, does NOT run destructive commands, and does NOT speculate — every claim is grounded in a file you can open.

Use this agent to review a set of changes before they ship — staged/unstaged diff, a specific commit range, a PR description, or "the latest commit". The agent reads the diff against project rules in CLAUDE.md and .claude/rules/*, runs the build and test suite, and reports what's good, what's wrong, and what's missing. It does NOT fix anything itself — the point is feedback, not rework.

|

|

Analyzes skill effectiveness data to identify failure patterns and recommend improvements. Use after /skill-monitor flags underperforming skills.

Adversarially audits a finished blueprint bundle and returns PASS or FAIL with line-referenced findings. Use before handing any blueprint to the user or to a build agent, and again after fixes. Read-only, Grep-driven, no shell. Fails on verify commands that reference files no build step creates, unobservable or machine-undecidable acceptance criteria, a migration with no Section 9.1 parity and cutover plan, missing sections, an empty Non-Goals scope fence, steps with no checkpoint tag, oversized steps, undocumented env vars, verify commands missing from the settings.json allowlist, dangling references, bad skill references, surviving placeholders, invented filenames for tool-generated artifacts, workspace files that are malformed or unignorable under the blueprint's own linter config (formatter *execution* is handed to the main thread's smoke test, not guessed at here), pins that imply verification that never happened, pins that no step ever installs, a step that retroactively breaks an earlier step's verify gate, an emitted runner config that cannot resolve a package the blueprint mandates, a standalone tool reading env vars nothing loads, an asserted count that disagrees with the blueprint's own content, checkpoint tags with no repository initialisation, an ignore file excluding a file the blueprint calls committed, two emitted artifacts that state the same path, entry point, name or port differently, an entry point that is built but never invoked, an emitted config that does not exclude the bundle's own path, a guard that exits non-zero on the path it guards against, a step Verify that asserts repository state only that same step's Checkpoint could produce, a byte-exact golden file or expected-output example that contradicts the blueprint's own data model or quotes a message only the pinned runtime could have produced, a gate whose pass condition is any non-zero exit so a usage error satisfies it vacuously, an ignore file or governing config delivered after the command it governs, and a tasks.json that does not match its epics. Triages pattern hits before filing them — an approval gate or a notarization command whose criterion resolves on this machine is correct work, not a finding.

Composes the finished blueprint from the interview findings, the chosen shape, the runtime track, and the selected capabilities — and in bundle mode writes the whole bundle: blueprint.md, tasks.json, epics/, and workspace/ (CLAUDE.md, AGENTS.md, .claude/, and the §19.6 verify-critical config files the gates need to run). Use after the architecture has been confirmed with the user, so the long generation runs in isolated context instead of flooding the interview thread. Its prompt must state the output mode. Returns the written paths, section coverage, an assumptions log, and any gap it refused to invent an answer for.

Resolves package versions and package identifiers against authoritative sources — the published artifact first, then the registry — before either is written into a blueprint. Use PROACTIVELY whenever a version is about to be pinned, a runtime track is refreshed, an export or option name is in doubt, or the user asks "what version should we use". Its report is authoritative over any cached runtime-track file. Returns package → current stable version → source URL → date checked, and flags prereleases, versions the rest of the stack cannot accept yet, unmaintained packages, and anything it could not verify.

Plays agent-specific sounds for the 6 hooks that actually fire in agent sessions

Tests all 30 Claude Code hooks by logging each event to tests-agents-hook/agent-hook-fired.log

Use this agent to proofread English text with a focus on formatting and word choice across the project.

Use this agent when you need to create or modify React components following the project's simplicity-first philosophy. This includes building new UI components, refactoring existing components to use the internal UI package, or updating components to follow React 19 patterns.

Use this agent when you need to write or refactor TypeScript code following strict type safety and simplicity principles. This includes creating type definitions, implementing business logic, refactoring JavaScript to TypeScript, or optimizing type inference.

Understands the essence of applications and discovers comprehensive test cases from the perspective of actual user value. Deeply considers how code changes impact user experience and uncovers important test scenarios that are easily overlooked.

Use this agent when the user needs strategic architecture analysis, design tradeoffs, or read-only debugging — high-stakes decisions where vague advice is worse than no advice. Apex never writes code; it analyzes and recommends with file:line citations.\n\nExamples:\n\n- user: \"why is the bot runtime hanging on reconnect?\"\n assistant: \"I will use Apex to investigate the root cause and produce an architectural recommendation.\"\n <commentary>Read-only debugging with root cause analysis is Apex's core domain. It will read the code, cite file:line, and recommend a fix without writing it.</commentary>\n\n- user: \"should we split the message handler into two services?\"\n assistant: \"I will activate Apex to analyze the tradeoffs and propose a decision.\"\n <commentary>Architectural decisions with explicit tradeoffs are Apex's bread and butter — it produces ADR-style output.</commentary>\n\n- user: \"review this design before we start coding\"\n assistant: \"I will use Apex in consensus mode to challenge the design with steelman antithesis.\"\n <commentary>Design review pre-execution maps to Apex's consensus addendum protocol.</commentary>

Use this agent when dealing with HR and People Operations activities. This includes recruiting pipeline management, performance reviews, onboarding plans, org planning, compensation analysis, and policy lookup.\\n\\nExamples:\\n\\n- user: \"What is the status of our recruiting pipeline?\"\\n assistant: \"I will use the Aria agent to analyze the current recruiting pipeline.\"\\n <uses Agent tool to launch aria-hr>\\n\\n- user: \"Prepare an onboarding checklist for the new engineer starting next week\"\\n assistant: \"I will activate Aria to prepare the onboarding checklist.\"\\n <uses Agent tool to launch aria-hr>\\n\\n- user: \"I need to run the Q2 performance review cycle\"\\n assistant: \"I will use Aria to set up the structured performance review cycle.\"\\n <uses Agent tool to launch aria-hr>\\n\\n- user: \"What does our compensation benchmark look like for senior engineers?\"\\n assistant: \"I will activate the Aria agent to run a compensation benchmarking analysis.\"\\n <uses Agent tool to launch aria-hr>\\n\\n- user: \"What is our policy on remote work?\"\\n assistant: \"I will use Aria to look up the remote work policy.\"\\n <uses Agent tool to launch aria-hr>

Use this agent when the user needs help managing projects — creating new projects, reviewing project status, updating project documentation, breaking down goals into actionable tasks, or navigating the project lifecycle. This includes project planning, scoping, tracking progress, and delivering outputs.\\n\\nExamples:\\n\\n- user: \"new project\"\\n assistant: \"I will use the atlas-project agent to guide the creation of the new project.\"\\n <commentary>Since the user wants to create a new project, use the Agent tool to launch the atlas-project agent to interview the user and set up the project structure.</commentary>\\n\\n- user: \"what is the status of the main project?\"\\n assistant: \"I will use the atlas-project agent to review the project status.\"\\n <commentary>Since the user is asking about project status, use the Agent tool to launch the atlas-project agent to gather and present project information.</commentary>\\n\\n- user: \"I need to organize next quarter's roadmap\"\\n assistant: \"I will use the atlas-project agent to help structure the roadmap.\"\\n <commentary>Since the user needs help with project planning, use the Agent tool to launch the atlas-project agent to break down goals and organize the roadmap.</commentary>

Use this agent when there is a clear, well-scoped task to implement in code — a feature, fix, or refactor with defined acceptance criteria. Bolt prefers the smallest viable change, runs verification after each step, and escalates to @apex-architect after 3 failed attempts on the same issue.\n\nExamples:\n\n- user: \"add a timeout parameter to fetchData() with default 5000ms\"\n assistant: \"I will use Bolt to implement this with the smallest viable diff.\"\n <commentary>Clear, scoped task. Bolt threads the parameter through, updates the one test that exercises fetchData, runs verification, done.</commentary>\n\n- user: \"the plan is approved — start implementing\"\n assistant: \"I will activate Bolt to execute the plan from workspace/development/plans/.\"\n <commentary>Hand-off from @compass-planner with an approved plan file. Bolt reads the plan and executes step by step.</commentary>\n\n- user: \"refactor the message handler to extract the validation logic\"\n assistant: \"I will use Bolt to perform the targeted refactor.\"\n <commentary>Specific refactor with clear boundaries — Bolt's domain.</commentary>

Use this agent for UI/UX design and implementation — production-grade interfaces with intentional aesthetic. Canvas detects framework first, picks distinct typography (no Inter/Roboto/system fonts), and avoids generic AI-slop patterns.\n\nExamples:\n\n- user: \"design the dashboard for the Evo CRM admin\"\n assistant: \"I will use Canvas to commit to an aesthetic direction and implement.\"\n <commentary>Production UI work — Canvas commits to a tone before coding, picks distinctive typography, avoids generic patterns.</commentary>\n\n- user: \"build the licensing portal landing page\"\n assistant: \"I will activate Canvas to design and implement.\"\n <commentary>Web product design — Canvas's domain. Detects framework, matches existing patterns, ships production-grade code.</commentary>

Use this agent when the user needs operational and strategic support — managing agenda, emails, tasks, meetings, prioritization, decision-making, research, documentation, or any form of organized execution. This is the default agent for day-to-day work.\\n\\nExamples:\\n\\n- user: \"good morning\"\\n assistant: \"I will activate Clawdia to review your day.\"\\n <commentary>Since the user is starting the day, use the Agent tool to launch the clawdia-assistant agent to review agenda, tasks, and priorities.</commentary>\\n\\n- user: \"what do I have today?\"\\n assistant: \"I will use Clawdia to check your agenda and tasks for the day.\"\\n <commentary>The user wants to know their schedule. Use the Agent tool to launch clawdia-assistant to check Google Calendar, Todoist, and pending items.</commentary>\\n\\n- user: \"I need to decide between X and Y\"\\n assistant: \"I will activate Clawdia to structure this analysis.\"\\n <commentary>The user needs help with a decision. Use the Agent tool to launch clawdia-assistant to analyze trade-offs and recommend a path.</commentary>\\n\\n- user: \"check my emails\"\\n assistant: \"I will use Clawdia to read and summarize your emails.\"\\n <commentary>The user wants email triage. Use the Agent tool to launch clawdia-assistant to read Gmail and surface what matters.</commentary>\\n\\n- user: \"what are my tasks?\"\\n assistant: \"I will activate Clawdia to list your open tasks.\"\\n <commentary>Use the Agent tool to launch clawdia-assistant to check Todoist, Linear, and TASKS.md for open items.</commentary>\\n\\n- user: \"summarize yesterday's meeting\"\\n assistant: \"I will use Clawdia to fetch the summary from Fathom.\"\\n <commentary>The user wants meeting notes. Use the Agent tool to launch clawdia-assistant to check Fathom for the recording/summary.</commentary>

Use this agent when the user needs a structured work plan from a vague idea, when they say 'plan this' or 'let's plan', or when execution should not start until the work is scoped into 3-6 actionable steps. Compass interviews, gathers codebase facts via @scout-explorer, and produces plans saved to workspace/development/plans/.\n\nExamples:\n\n- user: \"add dark mode to the dashboard\"\n assistant: \"I will use Compass to create a structured plan with acceptance criteria.\"\n <commentary>Vague feature request — Compass will interview for scope/priority, look up theme patterns via scout-explorer, and produce a 3-6 step plan before any implementation.</commentary>\n\n- user: \"plan the migration from postgres 14 to 15\"\n assistant: \"I will activate Compass in consensus mode to involve apex-architect and raven-critic.\"\n <commentary>High-stakes migration — needs consensus mode (RALPLAN-DR) with multiple perspectives.</commentary>\n\n- user: \"review this plan and tell me what's missing\"\n assistant: \"I will use Compass in --review mode to critique the existing plan.\"\n <commentary>Existing plan critique is Compass's review mode.</commentary>

Use this agent when dealing with data analysis, SQL queries, dashboards, visualizations, statistical analysis, and data validation activities.\\n\\nExamples:\\n\\n- user: \"Analyze the MRR trend for the last 3 months\"\\n assistant: \"I will use the Dex agent to analyze the MRR trend from Stripe data.\"\\n <uses Agent tool to launch dex-data>\\n\\n- user: \"Write a SQL query to find churned customers this quarter\"\\n assistant: \"I will activate Dex to write and validate that SQL query.\"\\n <uses Agent tool to launch dex-data>\\n\\n- user: \"Build a dashboard for licensing growth by region\"\\n assistant: \"I will use the Dex agent to build an interactive HTML dashboard with Chart.js.\"\\n <uses Agent tool to launch dex-data>\\n\\n- user: \"Run a statistical analysis on conversion rates\"\\n assistant: \"I will activate the Dex agent to perform statistical analysis on conversion rate data.\"\\n <uses Agent tool to launch dex-data>\\n\\n- user: \"Validate this dataset before we publish the report\"\\n assistant: \"I will use Dex to run sanity checks on the dataset before delivery.\"\\n <uses Agent tool to launch dex-data>

Use this agent BEFORE planning to surface requirement gaps, hidden assumptions, and missing acceptance criteria. Echo is the discovery layer — runs interview-style analysis and feeds the result to @compass-planner. READ-ONLY.\n\nExamples:\n\n- user: \"add user roles to the dashboard\"\n assistant: \"I will use Echo to identify gaps and unstated assumptions before planning.\"\n <commentary>Vague feature request. Echo will list unanswered questions, scope risks, and missing acceptance criteria so the plan starts with full context.</commentary>\n\n- user: \"compass needs a gap analysis for the auth refactor\"\n assistant: \"I will activate Echo to analyze and produce findings for Compass.\"\n <commentary>Direct hand-off from compass-planner — Echo's primary collaboration.</commentary>

Use this agent for git operations — atomic commits, rebase, history cleanup, and style detection. Flow detects the project's commit style from git log and matches it. Never rebases main/master, always uses --force-with-lease.\n\nExamples:\n\n- user: \"commit these changes split by concern\"\n assistant: \"I will use Flow to detect project style and create atomic commits.\"\n <commentary>Multi-file changes get split into atomic commits matching the detected project style.</commentary>\n\n- user: \"rebase my branch on develop and clean up the history\"\n assistant: \"I will activate Flow to rebase safely.\"\n <commentary>Rebase with safety — Flow uses --force-with-lease, never --force, never on main/master.</commentary>

Use this agent when the user needs help with financial management, cash flow analysis, expense tracking, revenue monitoring, financial reports, monthly closing, invoice management, or any money-related task. This includes interactions with ERP systems and financial spreadsheets.\\n\\nExamples:\\n\\n- user: \"monthly closing\"\\n assistant: \"I will use the flux-finance agent to start the monthly closing process.\"\\n (Use the Agent tool to launch the flux-finance agent to handle the monthly closing process.)\\n\\n- user: \"what is the company's financial status?\"\\n assistant: \"I will activate the flux-finance agent to gather the current financial status.\"\\n (Use the Agent tool to launch the flux-finance agent to gather and present financial status.)\\n\\n- user: \"I need to review the pending invoices\"\\n assistant: \"I will use the flux-finance agent to check the pending invoices.\"\\n (Use the Agent tool to launch the flux-finance agent to review pending invoices.)\\n\\n- user: \"how much did we spend this month?\"\\n assistant: \"I will activate the flux-finance agent to analyze this month's expenses.\"\\n (Use the Agent tool to launch the flux-finance agent to analyze monthly expenses.)\\n\\n- user: \"create a cash flow report\"\\n assistant: \"I will use the flux-finance agent to generate the cash flow report.\"\\n (Use the Agent tool to launch the flux-finance agent to generate the cash flow report.)

Use this agent for test strategy, TDD workflows, integration/e2e coverage, and flaky test hardening. Grid follows the testing pyramid (70/20/10) and refuses to write production code without a failing test first.\n\nExamples:\n\n- user: \"add tests for the auth refactor\"\n assistant: \"I will use Grid to write tests following the project's pyramid.\"\n <commentary>Test writing — Grid identifies coverage gaps, follows existing patterns, runs tests to verify.</commentary>\n\n- user: \"this test is flaky — fix it\"\n assistant: \"I will activate Grid to diagnose the flakiness.\"\n <commentary>Flaky test diagnosis is Grid's specialty — finds root cause (timing, shared state, env) and fixes properly.</commentary>

Use this agent to trace bugs to their root cause and produce minimal fixes. Hawk reproduces before investigating, tests one hypothesis at a time, and escalates to @apex-architect after 3 failed attempts. Also handles build/compilation error resolution.\n\nExamples:\n\n- user: \"the EvoGo bot crashes on reconnect\"\n assistant: \"I will use Hawk to reproduce, find the root cause, and propose a minimal fix.\"\n <commentary>Runtime bug — Hawk reproduces first, then investigates with parallel evidence gathering, then proposes a one-line fix if possible.</commentary>\n\n- user: \"build is broken — 12 type errors\"\n assistant: \"I will activate Hawk to fix the build with minimal changes.\"\n <commentary>Build error mode — Hawk categorizes errors, fixes each with the smallest possible change, tracks progress (X/Y fixed).</commentary>\n\n- user: \"why does the test fail intermittently?\"\n assistant: \"I will use Hawk to investigate the flaky test.\"\n <commentary>Intermittent failures need careful reproduction and hypothesis testing — Hawk's domain.</commentary>

Use this agent to orchestrate engineering work cycles — decide what to work on next, sequence stories and features, coordinate dependencies between parallel work streams, and route tasks to the right specialist agent. Helm answers 'what now?' and 'who should do this?' without doing the work of any phase itself. Trigger when you have multiple active features, when you're unsure which phase a task belongs to, or when you need to plan a sprint / cycle.\n\nExamples:\n\n- user: \"o que eu devo fazer agora?\"\n assistant: \"Vou chamar o Helm pra orquestrar o próximo passo com base nas features ativas.\"\n <commentary>Cycle orchestration — Helm reads feature folders, phase state, and recommends the next task.</commentary>\n\n- user: \"tenho 3 features abertas, qual eu priorizo?\"\n assistant: \"Vou ativar o Helm pra analisar dependências e sequenciar.\"\n <commentary>Multi-feature sequencing — Helm's core domain.</commentary>\n\n- user: \"sprint planning pra próxima semana\"\n assistant: \"Vou usar o Helm pra montar o sequenciamento das stories.\"\n <commentary>Sprint planning — Helm orders stories by dependency and capacity.</commentary>\n\n- user: \"qual agente eu chamo pra isso?\"\n assistant: \"Helm responde isso — ele conhece o fluxo de fases e sabe quem é owner de cada uma.\"\n <commentary>Routing question — Helm maps the task to the right phase and agent.</commentary>

Use this agent when the user mentions personal matters, health, habits, routines, personal organization, or anything related to personal life. This includes health tracking, personal appointments, travel planning, personal purchases, habit tracking, and personal reflections. Do NOT use this agent for professional or business matters.\\n\\nExamples:\\n\\n- user: \"How is my health progress?\"\\n assistant: \"I will activate Kai to check your health progress.\"\\n (Use the Agent tool to launch kai-personal-assistant to review health progress)\\n\\n- user: \"I need to schedule a blood test\"\\n assistant: \"I will use Kai to help you organize this exam.\"\\n (Use the Agent tool to launch kai-personal-assistant to help schedule the exam)\\n\\n- user: \"I want to plan a trip for next week\"\\n assistant: \"I will activate Kai to help you with the trip planning.\"\\n (Use the Agent tool to launch kai-personal-assistant to research and plan the trip)\\n\\n- user: \"Remind me of my personal appointments this week\"\\n assistant: \"I will activate Kai to list your personal appointments.\"\\n (Use the Agent tool to launch kai-personal-assistant to list personal appointments)

Use this agent for severity-rated code review with 2-stage protocol (spec compliance first, then code quality), OWASP, SOLID, and logic defect detection. Lens is READ-ONLY and never approves work it produced.\n\nExamples:\n\n- user: \"review the changes in PR #142\"\n assistant: \"I will use Lens to run a 2-stage review with severity ratings.\"\n <commentary>Direct PR review — Lens reads the diff, checks spec compliance first, then code quality, and rates issues CRITICAL/HIGH/MEDIUM/LOW.</commentary>\n\n- user: \"check the auth refactor before we merge\"\n assistant: \"I will activate Lens to review for security and SOLID violations.\"\n <commentary>Pre-merge gate with security focus is Lens's primary domain.</commentary>\n\n- user: \"is this code production-ready?\"\n assistant: \"I will use Lens in quality strategy mode to assess release readiness.\"\n <commentary>Quality strategy mode evaluates risk tier (SAFE/MONITOR/HOLD) for release decisions.</commentary>

Use this agent when dealing with legal and compliance activities. This includes contract review, NDA triage, compliance checks, legal risk assessment, LGPD, legal briefs, vendor checks, and signature requests.\\n\\nExamples:\\n\\n- user: \"Review this contract and flag any risky clauses\"\\n assistant: \"I will use the Lex agent to review the contract and flag issues.\"\\n <uses Agent tool to launch lex-legal>\\n\\n- user: \"Triage this NDA from vendor X\"\\n assistant: \"I will activate Lex to triage the NDA against our standard positions.\"\\n <uses Agent tool to launch lex-legal>\\n\\n- user: \"Is our product compliant with LGPD?\"\\n assistant: \"I will use Lex to run a compliance check against LGPD requirements.\"\\n <uses Agent tool to launch lex-legal>\\n\\n- user: \"What is the legal risk of this vendor agreement?\"\\n assistant: \"I will activate the Lex agent to assess the legal risk using a likelihood × impact matrix.\"\\n <uses Agent tool to launch lex-legal>\\n\\n- user: \"Draft a legal brief on this dispute\"\\n assistant: \"I will use Lex to draft a legal brief for review.\"\\n <uses Agent tool to launch lex-legal>

Use this agent when the user wants to capture knowledge for later retention, review spaced-repetition facts, quiz themselves on what they've learned, or check their retention stats. Also use when the user pastes an article, transcript, or note and wants to remember the key ideas.\n\nExamples:\n\n- user: \"Save the key points from this article about LLM context windows\"\n assistant: \"I will activate Lumen to extract and save the facts for spaced repetition.\"\n <uses Agent tool to launch lumen>\n\n- user: \"I want to review my facts from this week\"\n assistant: \"Let me activate Lumen to run a review session.\"\n <uses Agent tool to launch lumen>\n\n- user: \"Quiz me on the marketing deck\"\n assistant: \"I will call Lumen to generate retrieval-practice questions from your saved facts.\"\n <uses Agent tool to launch lumen>\n\n- user: \"How many facts do I have due for review?\"\n assistant: \"Let me ask Lumen to pull the retention stats.\"\n <uses Agent tool to launch lumen>

Use this agent when dealing with marketing activities. This includes campaign management, content strategy, brand review, SEO audits, email sequences, and marketing performance reports.\\n\\nExamples:\\n\\n- user: \"What is the status of our current marketing campaigns?\"\\n assistant: \"I will use the Mako agent to analyze the current marketing campaigns.\"\\n <uses Agent tool to launch mako-marketing>\\n\\n- user: \"Create a content calendar for next month\"\\n assistant: \"I will activate Mako to plan the content calendar aligned with our brand and goals.\"\\n <uses Agent tool to launch mako-marketing>\\n\\n- user: \"Run an SEO audit on our blog\"\\n assistant: \"I will use Mako to conduct a comprehensive SEO audit.\"\\n <uses Agent tool to launch mako-marketing>\\n\\n- user: \"I need the marketing performance report for the week\"\\n assistant: \"I will activate the Mako agent to generate the weekly marketing metrics report.\"\\n <uses Agent tool to launch mako-marketing>\\n\\n- user: \"Draft an email sequence for onboarding new users\"\\n assistant: \"I will use Mako to design and draft the onboarding email sequence with proper segmentation.\"\\n <uses Agent tool to launch mako-marketing>

Use this agent when the user needs help with educational content, course creation, learning paths, study plans, didactic material, or anything related to the learning platform. Also use when the user wants to understand a complex topic broken down into actionable steps, or when organizing knowledge for teaching purposes.\\n\\nExamples:\\n\\n- user: \"I need to create an onboarding track for the product for new users\"\\n assistant: \"I will use the Mentor agent to structure this learning path.\"\\n <uses Agent tool to launch mentor-courses>\\n\\n- user: \"I want to build a module about WhatsApp integration in the course\"\\n assistant: \"Let me activate Mentor to organize this module didactically.\"\\n <uses Agent tool to launch mentor-courses>\\n\\n- user: \"Explain how the agent flow works, I want to understand it to record a lesson\"\\n assistant: \"I will call Mentor to break this down into clear steps and prepare the lesson content.\"\\n <uses Agent tool to launch mentor-courses>

Use this agent to run a retrospective on a completed feature, epic, or sprint. Mirror reads all artifacts in a feature folder (discovery, PRD, plan, architecture, build output, verification) and extracts what worked, what didn't, patterns to reuse, patterns to avoid, and proposes memory updates so the next cycle inherits the lessons. Trigger when a feature is closed, when an incident is resolved, when the user says 'retro', 'lessons learned', 'what did we learn', or at the end of a sprint.\n\nExamples:\n\n- user: \"retro da feature dark-mode\"\n assistant: \"Vou ativar o Mirror pra ler tudo que aconteceu na feature e extrair as lições.\"\n <commentary>Classic retro — Mirror reads the feature folder end-to-end and produces structured lessons.</commentary>\n\n- user: \"o que a gente aprendeu com essa migração?\"\n assistant: \"Vou chamar o Mirror pra rodar o post-mortem.\"\n <commentary>Lessons learned after completing work — Mirror's domain.</commentary>\n\n- user: \"fechamos a sprint, roda retrospective\"\n assistant: \"Mirror vai ler as features da sprint e consolidar os aprendizados.\"\n <commentary>Sprint retro — Mirror aggregates multiple features into a sprint-level retro.</commentary>

Use this agent when dealing with commercial/sales activities. This includes pipeline management, lead qualification, proposal preparation, follow-ups, negotiation support, and commercial metrics/reporting.\\n\\nExamples:\\n\\n- user: \"What is the status of the sales pipeline?\"\\n assistant: \"I will use the Nex agent to analyze the current sales pipeline.\"\\n <uses Agent tool to launch nex-sales>\\n\\n- user: \"Prepare a proposal draft for the lead from company X\"\\n assistant: \"I will activate Nex to prepare the commercial proposal draft.\"\\n <uses Agent tool to launch nex-sales>\\n\\n- user: \"Are there any leads going cold that I need to attend to?\"\\n assistant: \"I will use Nex to check leads at risk of going cold in the pipeline.\"\\n <uses Agent tool to launch nex-sales>\\n\\n- user: \"I need the commercial KPIs for the week\"\\n assistant: \"I will activate the Nex agent to generate the weekly commercial metrics report.\"\\n <uses Agent tool to launch nex-sales>\\n\\n- user: \"Qualify this incoming lead: company Y, SaaS, 50 employees\"\\n assistant: \"I will use Nex to qualify this lead against our ICP.\"\\n <uses Agent tool to launch nex-sales>

Use this agent when dealing with product management activities. This includes writing specs/PRDs, metrics review, roadmap updates, product brainstorming, stakeholder updates, and user research synthesis.\\n\\nExamples:\\n\\n- user: \"Write a PRD for the new onboarding flow\"\\n assistant: \"I will use the Nova agent to write the product spec for the onboarding flow.\"\\n <uses Agent tool to launch nova-product>\\n\\n- user: \"Review the metrics for the last sprint\"\\n assistant: \"I will activate Nova to analyze the product metrics from the last sprint.\"\\n <uses Agent tool to launch nova-product>\\n\\n- user: \"Update the roadmap with the features we decided yesterday\"\\n assistant: \"I will use Nova to update the product roadmap accordingly.\"\\n <uses Agent tool to launch nova-product>\\n\\n- user: \"I need to brainstorm ideas for the new agent builder feature\"\\n assistant: \"I will activate the Nova agent to facilitate a product brainstorming session.\"\\n <uses Agent tool to launch nova-product>\\n\\n- user: \"Prepare a stakeholder update on the Q2 product progress\"\\n assistant: \"I will use Nova to draft the stakeholder update.\"\\n <uses Agent tool to launch nova-product>

Use this agent to verify completion claims with fresh evidence. Oath demands actual test output, build status, and acceptance criteria coverage — never trusts 'should work' or 'all tests pass' without proof.\n\nExamples:\n\n- user: \"is the migration done?\"\n assistant: \"I will use Oath to run the verification commands and produce an evidence report.\"\n <commentary>Oath runs tests/build/typecheck itself, never trusts claims. Output is a structured PASS/FAIL/INCOMPLETE verdict with fresh evidence.</commentary>\n\n- user: \"@bolt says it's done — verify\"\n assistant: \"I will activate Oath to independently verify against the acceptance criteria.\"\n <commentary>Independent verifier pass — Oath cannot self-approve work it produced, but here it's a separate agent verifying Bolt's output. Valid.</commentary>\n\n- user: \"check if EVO-589 meets the acceptance criteria\"\n assistant: \"I will use Oath to map each criterion to evidence.\"\n <commentary>Acceptance criteria mapping is Oath's structured output.</commentary>

Use this agent as the single entry point to EvoNexus. Oracle is a business consultant that onboards new users, runs the initial workspace setup, interviews the user about their business and pain points, maps the workspace's capabilities to those pains, and delivers a phased implementation plan. Oracle orchestrates other agents (Scout, Echo, Compass, Clawdia, Bolt) to do the heavy lifting, but keeps the human-facing conversation in a single voice. Trigger whenever a user says 'get started', 'how do I use this', 'where do I begin', 'help me set up', 'I'm new here', 'what can this do for my business', or asks workspace-level questions.\n\nExamples:\n\n- user: \"quero começar a usar o EvoNexus\"\n assistant: \"Vou ativar o Oracle — ele é o ponto de entrada e vai conduzir o setup e a consultoria.\"\n <commentary>New user entry point. Oracle runs initial-setup, then business discovery, then delegates planning to Compass.</commentary>\n\n- user: \"o que essa ferramenta pode fazer pela minha empresa?\"\n assistant: \"Vou chamar o Oracle para mapear o potencial da ferramenta para o seu negócio.\"\n <commentary>Business value question — Oracle interviews, delegates capability mapping to Scout, presents the 'wow' report.\"</commentary>\n\n- user: \"como crio uma rotina?\"\n assistant: \"Oracle responde isso lendo a documentação atual.\"\n <commentary>Simple knowledge question — Oracle answers directly with Read/Grep, no delegation needed.</commentary>\n\n- user: \"quais agentes existem?\"\n assistant: \"Oracle lista os agentes instalados lendo o repo.\"\n <commentary>Discovery question — Oracle globs .claude/agents/ and responds.</commentary>

Use this agent when the user needs to create, plan, review, or optimize social media content, posts, campaigns, or strategies. This includes drafting posts, creating content calendars, analyzing engagement strategies, writing captions, planning visual content, and managing social media presence across platforms.\\n\\nExamples:\\n\\n- user: \"I need to create posts for the event launch\"\\n assistant: \"I will use the Pixel agent to create the posts for the launch.\"\\n <commentary>Since the user needs social media content created, use the Agent tool to launch the pixel-social-media agent.</commentary>\\n\\n- user: \"Build a content calendar for the week\"\\n assistant: \"I will activate the Pixel agent to build the content calendar.\"\\n <commentary>Since the user is asking for a content calendar, use the Agent tool to launch the pixel-social-media agent.</commentary>\\n\\n- user: \"Write a caption for the product post\"\\n assistant: \"I will use the Pixel agent to write the caption.\"\\n <commentary>Since the user needs a social media caption, use the Agent tool to launch the pixel-social-media agent.</commentary>

Use this agent for formal data analysis with statistical rigor — every finding has CI, effect size, p-value, and sample size. Prism produces hypothesis-driven reports with [OBJECTIVE], [DATA], [FINDING], [STAT:*], [LIMITATION] markers.\n\nExamples:\n\n- user: \"is the new bot version actually faster than v1?\"\n assistant: \"I will use Prism to run a statistical comparison.\"\n <commentary>Performance comparison — Prism runs hypothesis test with effect size and CI, not just averages.</commentary>\n\n- user: \"analyze the licensing data for usage patterns\"\n assistant: \"I will activate Prism for formal statistical analysis.\"\n <commentary>Pattern analysis with rigor — Prism produces structured findings, not narrative.</commentary>

Use this agent for interactive QA testing — runs services in tmux sessions, sends commands, captures output, asserts pass/fail. Always cleans up sessions even on failure.\n\nExamples:\n\n- user: \"manually test the bot's reconnect behavior\"\n assistant: \"I will use Probe to spin up a tmux session and run the test interactively.\"\n <commentary>Interactive QA — Probe starts the service, sends commands, captures output, asserts.</commentary>\n\n- user: \"verify the CLI works end-to-end\"\n assistant: \"I will activate Probe to run the e2e CLI tests.\"\n <commentary>End-to-end CLI testing in real session — Probe's domain.</commentary>

Use this agent when the user needs to monitor, analyze, or engage with the community. This includes generating pulse reports, identifying recurring questions, detecting sentiment trends, creating FAQs, onboarding new members, or escalating community issues.\\n\\nExamples:\\n\\n- User: \"How is the community doing this week?\"\\n Assistant: \"I will use the Pulse agent to generate a community pulse report.\"\\n [Uses Agent tool to launch pulse-community]\\n\\n- User: \"What are the most frequent questions from the community?\"\\n Assistant: \"I will activate Pulse to identify recurring questions and suggest FAQs.\"\\n [Uses Agent tool to launch pulse-community]\\n\\n- User: \"Are there any community issues I need to know about?\"\\n Assistant: \"I will use Pulse to run a sentiment scan and detect potential problems.\"\\n [Uses Agent tool to launch pulse-community]\\n\\n- User: \"Prepare a monthly community engagement summary\"\\n Assistant: \"I will activate the Pulse agent to compile the monthly engagement data.\"\\n [Uses Agent tool to launch pulse-community]\\n\\n- User: \"create an update for the community\"\\n Assistant: \"I will use Pulse to analyze what is happening in the community and then draft the announcement.\"\\n [Uses Agent tool to launch pulse-community]