produtoramaxvision
UserDynamic skill and subagent orchestrator for Claude Code. Lazy-load 500+ specialist agents and 4,000+ skills via FTS5/BM25 catalog from 20 curated sources. MIT-licensed.
Categories
Indexed Skills (74)
analyze-candidates
Ranks candidate skills/agents by task fit using Sonnet LLM-as-judge AND classifies task complexity (model + effort) in same call. Input is union of cheatsheet + FTS5 candidates with EQUAL weight per D3+D9. Output is task-driven ranking + complexity classification + INSTALL_BEST | INSTALL_MULTIPLE | SYNTHESIZE recommendation. v0.5.0 replaces v0.4.0 BM25+Jaccard ranking with Sonnet-as-judge (Jaccard retained only for INSTALL_BEST→MULTIPLE downgrade cross-check).
cso
Chief Security Officer mode. Infrastructure-first security audit: secrets archaeology, dependency supply chain, CI/CD pipeline security, LLM/AI security, skill supply chain scanning, plus OWASP Top 10, STRIDE threat modeling, and active verification. Two modes: daily (zero-noise, 8/10 confidence gate) and comprehensive (monthly deep scan, 2/10 bar). Trend tracking across audit runs. Use when: "security audit", "threat model", "pentest review", "OWASP", "CSO review". (MaxVision) Voice triggers (speech-to-text aliases): "see-so", "see so", "security review", "security check", "vulnerability scan", "run security".
design-consultation
Design consultation: understands your product, researches the landscape, proposes a complete design system (aesthetic, typography, color, layout, spacing, motion), and generates font+color preview pages. Creates DESIGN.md as your project's design source of truth. For existing sites, use /plan-design-review to infer the system instead. Use when asked to "design system", "brand guidelines", or "create DESIGN.md". Proactively suggest when starting a new project's UI with no existing design system or DESIGN.md. (MaxVision)
design-html
Design finalization: generates production-quality Pretext-native HTML/CSS. Works with approved mockups from /design-shotgun, CEO plans from /plan-ceo-review, design review context from /plan-design-review, or from scratch with a user description. Text actually reflows, heights are computed, layouts are dynamic. 30KB overhead, zero deps. Smart API routing: picks the right Pretext patterns for each design type. Use when: "finalize this design", "turn this into HTML", "build me a page", "implement this design", or after any planning skill. Proactively suggest when user has approved a design or has a plan ready. (MaxVision) Voice triggers (speech-to-text aliases): "build the design", "code the mockup", "make it real".
design-review
Designer's eye QA: finds visual inconsistency, spacing issues, hierarchy problems, AI slop patterns, and slow interactions — then fixes them. Iteratively fixes issues in source code, committing each fix atomically and re-verifying with before/after screenshots. For plan-mode design review (before implementation), use /plan-design-review. Use when asked to "audit the design", "visual QA", "check if it looks good", or "design polish". Proactively suggest when the user mentions visual inconsistencies or wants to polish the look of a live site. (MaxVision)
design-shotgun
Design shotgun: generate multiple AI design variants, open a comparison board, collect structured feedback, and iterate. Standalone design exploration you can run anytime. Use when: "explore designs", "show me options", "design variants", "visual brainstorm", or "I don't like how this looks". Proactively suggest when the user describes a UI feature but hasn't seen what it could look like. (MaxVision)
detect-multi-task
Detects whether a user task contains multiple independent subtasks and splits it into a JSON array. Internal helper for orchestrate Phase 0.5; uses session credentials (no API key). Always returns STRICT JSON {is_multi, tasks[]} with cap N=5.
dispatch-background
Dispatches a long-running task to a maxvision-* subagent in background via Claude Code's Agent View (`claude --bg`). Routes the task through `route-task` to pick the right agent, then forwards to `bin/maxvision-dispatch run`. Use when the task is long-running and you want to keep the main session unblocked. The session appears as a row in `claude agents`.
execute-phase-bg
Background-dispatch variant of execute-phase. Spawns one `claude --bg --agent maxvision-executor` session per plan (with auto-worktree) instead of running waves of in-process Agent calls. Frees the main session immediately. Use when the phase has multiple independent plans, the work is long-running, and you want to monitor via `claude agents` instead of staying attached.
keyword-extractor
Extract 3-7 technical English keywords from a routing query. Internal helper for route-task Stage 1; uses session credentials (no API key). Translates pt-BR/ES/ZH to EN, ignores stopwords, keeps technical terms verbatim (FTS5, sqlite3, react, B3, B4).
brainstorming
Use BEFORE any creative work — features, components, behavior changes. Explores intent + requirements + design before implementation.
check-update
Owner-only — checks upstream versions of maxvision plus the 3 upstream sources (maxvision, upstream, maxvision). Gated by gh CLI auth.
check-version
Compares the locally installed version of a skill or subagent against its upstream source (sha or semver) and reports whether an update is available. Cache-aware (1-hour TTL). Used by `orchestrate` before dispatching a subagent or invoking a skill, to guarantee the components are fresh.
cleanup-ephemerals
Sweeps both ~/.claude/skills/*/.maxvision-source.json AND ~/.claude/agents/.maxvision-source-*.json for ephemeral=true entries at SessionEnd. Deletes those below promote_threshold (default 5 uses); promotes skills at/above threshold by rewriting ephemeral=false. Invoked automatically by the SessionEnd hook unless reason=resume. Backward compat — v1.0.0 skill manifests with no ephemeral field stay permanent; v1.0.0 agent sidecars with no ephemeral field are auto-migrated to ephemeral=true (WR-04, v1.1.7+).
discover-agent
Looks up subagents by keyword/domain across all installed plugins (voltagent, pr-review-toolkit, feature-dev, code-review, maxvision, etc.) plus user-level agents. Twin of discover-skill but searches kind='agent' in the FTS5 index. Cross-references with CLAUDE-agents.md cheatsheet for trusted-mappings.
discover-skill
Looks up a skill by name, keyword, or domain in the curated catalog (skill-sources.json). Returns the source repo, tier of trust, install methods supported, and any guardrails (overlap warnings, deprecations). Use when the orchestrator detects a missing skill and needs to know where to fetch it from.
executing-plans
Use when you have a written implementation plan to execute in a separate session with review checkpoints
grill-me
Clarity gate that refines vague or ambiguous tasks via short Socratic dialogue before keyword extraction and routing. Asks 2-4 focused, one-at-a-time questions targeting scope, success criteria, target files, constraints, and expected output format. Returns a REFINED_TASK string that orchestrate uses for downstream phases. Called from orchestrate P-1; can also be invoked directly when a request feels under-specified.
grill-with-docs
DDD doc gate that writes or updates CONTEXT.md and appends an Architecture Decision Record (ADR) to docs/adr/ when the refined task targets a writer subagent in a repo that already shows DDD signal. Sharpens fuzzy domain terminology against the existing glossary before the writer dispatches. Called from orchestrate P7.4 after Phase 7.5 provisioning; no-op when the gate conditions are not met.
index-catalog
Bootstraps or refreshes the local FTS5 catalog at ~/.claude/cache/maxvision/index.db by downloading the latest release artifact (default) or rebuilding it locally with --rebuild. Use after first install OR when version-check shows the index is stale OR when you want to force a refresh.
install-agent
Installs a Claude Code subagent on-demand from the FTS5 catalog (`~/.claude/cache/maxvision/index.db`) into `~/.claude/agents/<name>.md` (user scope) or `.claude/agents/<name>.md` (project scope). Twin of `install-skill` for agents. Writes `.maxvision-source-<name>.json` sidecar for version tracking. Used by `orchestrate` Phase 7.3 when an agent is needed but not present locally, enabling lazy-load model that bypasses the cumulative-description token budget.
install-skill
Installs a Claude Code skill from a curated source via one of three methods (sparse-clone, plugin marketplace, or npx antigravity filtered). Always asks for confirmation before any install. Creates `.maxvision-source.json` in the skill folder for traceability. Used by `orchestrate` after `discover-skill` identifies missing components. v0.2.0 adds a 4th method `synthesis` that delegates to `synthesize-skill` for SYNTHESIZE-mode recommendations from `analyze-candidates`.
autoplan
Orchestrate the four post-plan specialist reviewers (CEO + Eng + Design + DevEx) and consolidate findings. Use after writing-plans to auto-run all reviewer skills.
canary
Canary release verification — sample traffic + metric guards + auto-rollback. Use after deploy to validate before full traffic exposure.
document-generate
Generate user-facing documentation using Diataxis quadrants (tutorial/how-to/reference/explanation). Use when documenting a new feature surface.
document-release
Generate release notes, changelog entry, and migration guidance from a merged PR set. Use after squash-merge to publish a v<X> release narrative.
freeze
Freeze a module so future edits trigger advisory warnings until explicitly unfrozen. Use to mark a module post-incident as do-not-touch without review.
guard
Combined careful + freeze in one skill: stricter gates plus module freeze in one operation. Use when both layers should fire on the same target.
careful
Enter careful-mode: stricter pre-edit gates and auto-checkpointing for high-stakes work. Use before changes to load-bearing or security-critical code.
add-tests
Generate tests from UAT criteria respecting language conventions and test-driven-development doctrine. Use to backfill or harden test coverage.
ai-integration-phase
Complete AI integration phase — selects framework, researches implementation, audits domain context, and designs evaluation strategy. Orchestrates maxvision-framework-selector → maxvision-ai-researcher → maxvision-domain-researcher → maxvision-eval-planner. Produces AI-SPEC.md. Use before plan-phase for any AI/LLM feature.
code-review
Bug/security/quality review of source + automated fix application, consolidated. Use after a phase or PR diff to catch regressions before ship.
complete-milestone
Close a completed milestone. Audits open artifacts, verifies all phases are VERIFIED, archives to .maxvision/milestones/v{X.Y}/, updates MILESTONES.md and ROADMAP.md, creates git tag. Gate — all phases must have SUMMARY.md before closing.
context-restore
Restore a previously saved working context from a /context-save checkpoint. Loads the most recent save by default (cross-branch). Supports /context-restore <title-fragment|number>.
context-save
Save current conversation context as a WIP commit with [maxvision-context] body to restore later. Use /context-restore to resume. Also supports: /context-save list [--all]
debug
Persistent-state phase-cycle debug — hypothesis tracking + iterative reproduction. Use when bug needs multi-session memory or is non-deterministic.
discuss-phase
Extract implementation gray areas + capture decisions via AskUserQuestion modes. Use when scoping a phase before planning, or to lock unknowns.
dispatching-parallel-agents
Use when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies
eval-review
Retroactive audit of an implemented AI phase's evaluation coverage. 2 states — A (AI-SPEC.md + SUMMARY exists, audit against spec) B (SUMMARY only, audit against best practices). Spawns maxvision-eval-auditor. Produces scored EVAL-REVIEW.md with verdict, score/100, and critical gap count.
execute-phase
Drive execution of a planned phase with codebase-drift, per-plan-worktree and post-merge gates. Use when starting implementation after plan approval.
execute-plan
Run a single PLAN.md inside a phase, atomic commits per task with deviation handling. Use when one plan inside a multi-plan phase is ready.
extract-learnings
Extract institutional knowledge from completed phase artifacts into structured LEARNINGS.md. 4 categories — decisions, lessons, patterns, surprises. Requires PLAN.md + SUMMARY.md; optional VERIFICATION.md, UAT.md, STATE.md. Overwrites previous LEARNINGS.md. No subagents — orchestrator reads and writes directly.
fast
Inline trivial fixes + skip-optional-agents mode, consolidated. Use when work is small enough that ceremony costs more than the change itself.
finishing-a-development-branch
Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup
forensics
Post-mortem analysis of MaxVision workflow failures. Read-only investigation of git history and .maxvision/ artifacts to detect stuck loops, missing artifacts, abandoned work, crashes, scope drift, and test regressions. Writes structured report to .maxvision/forensics/.
ingest-docs
Bootstrap or merge .maxvision/ project context from existing ADRs, PRDs, SPECs, and DOCs. Classifies documents in parallel, synthesizes with conflict detection, and routes to new-project or existing roadmap. Use when onboarding a brownfield project with existing documentation.
investigate
Pre-fix Iron Law: never patch a bug without investigating root cause first; auto-freeze the suspected module. Use the moment a bug surfaces, before any edit.
capture
Codex-native MaxVision capture entrypoint for ideas, tasks, notes, backlog items, seeds, and interactive lists.
cleanup
Codex-native MaxVision cleanup entrypoint for archiving completed phase directories and related project artifacts.
enable-live-countdown
Codex compatibility entrypoint for MaxVision live countdown/statusline setup without changing Claude Code settings.
health
Codex-native MaxVision health check for planning directory state, repair candidates, and workflow consistency.
inbox
Codex-native MaxVision inbox triage for GitHub issues and pull requests against project templates.
capability-matrix
Model x params capability cheatsheet. Static reference; read by all media-forge agents.
media-forgehiggsfield-prompting
Higgsfield prompting playbook — MCSLA formula, DoP camera verbs, Cinema Studio lens dictionary, Soul ID lifecycle, Marketing Studio templates. Trigger when caller plans a Higgsfield generation.
audit
Audit a job's lineage: read metadata, trace, lineage, verdicts. Triggers: audit, lineage, verdict, history.
brand-check
INTERNAL - brand compliance check (color deltaE2000 + logo + font). Used by enterprise-corrector and quality-reviewer.
campaign
Multi-asset campaign: generates a coordinated set (hero + variations + sizes). Triggers: campaign, ad campaign, multi-asset, batch.
character-sheet
Character consistency: identity-lock portrait + turnaround + expression sheet from one reference. Triggers: character sheet, turnaround, consistency.
cinematic-short
Cinematic short video workflow: storyboard -> frame imagery -> video chain. Triggers: cinematic short, narrative video, short film.
create
Main orchestrator. Refines user brief, routes to domain agent, runs reviewer, surfaces verdict. Triggers: create, generate, make, produce.
extend-video
Extend existing video via Veo 3.1 +7s hops. Triggers: extend video, longer video, continuation.
ocr-validate
INTERNAL - OCR text validation via Cloud Vision. Used by quality-reviewer.
scene-compose
Multi-image scene composition with role-labeled refs (up to 14). Triggers: scene compose, composition, multi-image scene.
setup
Onboarding wizard: env detection, API key prompts, output dir setup, first dry-run smoke. Run on first install.
kling-prompting
Kling V3 (Standard / Pro / Master 4K / Omni multi-shot) prompt engineering. 5-part spine (Camera-Scene-Action-Vibe-Time), per-mode cookbook, model-tier selection. Use BEFORE invoking media_kling_* tools or kling-director.
media-forgeseedance-prompting
Seedance 2.0 prompt engineering: tier selection (Fast 720p / Standard 1080p — no Pro tier), multi-shot timestamp markers, @-mention reference syntax for reference-to-video, audio+video joint generation, image-to-video end-frame transitions. Use BEFORE invoking media_seedance_* tools or seedance-director.
gtm-api
Execute Google Tag Manager API operations - create, update, delete, and manage tags, triggers, variables, and containers. Use when working with GTM programmatically, generating API calls, validating GTM configurations, publishing container versions, or automating tag management workflows.
n8n-code-javascript
Write JavaScript code in n8n Code nodes. Use when writing JavaScript in n8n, using $input/$json/$node syntax, making HTTP requests with $helpers, working with dates using DateTime, troubleshooting Code node errors, choosing between Code node modes, or doing any custom data transformation in n8n. Always use this skill when a workflow needs a Code node — whether for data aggregation, filtering, API calls, format conversion, batch processing logic, or any custom JavaScript. Covers SplitInBatches loop patterns, cross-iteration data, pairedItem, and real-world production patterns.
n8n-code-python
Write Python code in n8n Code nodes. Use when writing Python in n8n, using _input/_json/_node syntax, working with standard library, or need to understand Python limitations in n8n Code nodes. Use this skill when the user specifically requests Python for an n8n Code node. Note — JavaScript is recommended for 95% of use cases — only use Python when the user explicitly prefers it or the task requires Python-specific standard library capabilities (regex, hashlib, statistics).
n8n-expression-syntax
Validate n8n expression syntax and fix common errors. Use when writing n8n expressions, using {{}} syntax, accessing $json/$node variables, troubleshooting expression errors, mapping data between nodes, or referencing webhook data in workflows. Use this skill whenever configuring node fields that reference data from previous nodes — expressions are how n8n passes data between nodes, and getting the syntax wrong is the most common source of workflow errors.
n8n-mcp-tools-expert
Expert guide for using n8n-mcp MCP tools effectively. Use when searching for nodes, validating configurations, accessing templates, managing workflows, managing credentials, auditing instance security, or using any n8n-mcp tool. Provides tool selection guidance, parameter formats, and common patterns. IMPORTANT — Always consult this skill before calling any n8n-mcp tool — it prevents common mistakes like wrong nodeType formats, incorrect parameter structures, and inefficient tool usage. If the user mentions n8n, workflows, nodes, or automation and you have n8n MCP tools available, use this skill first.
n8n-node-configuration
Operation-aware node configuration guidance. Use when configuring nodes, understanding property dependencies, determining required fields, choosing between get_node detail levels, or learning common configuration patterns by node type. Always use this skill when setting up node parameters — it explains which fields are required for each operation, how displayOptions control field visibility, and when to use patchNodeField for surgical edits vs full node updates.
n8n-validation-expert
Interpret validation errors and guide fixing them. Use when encountering validation errors, validation warnings, false positives, operator structure issues, or need help understanding validation results. Also use when asking about validation profiles, error types, the validation loop process, or auto-fix capabilities. Consult this skill whenever a validate_node or validate_workflow call returns errors or warnings — it knows which warnings are false positives and which errors need real fixes.
n8n-workflow-patterns
Proven workflow architectural patterns from real n8n workflows. Use when building new workflows, designing workflow structure, choosing workflow patterns, planning workflow architecture, or asking about webhook processing, HTTP API integration, database operations, AI agent workflows, batch processing, or scheduled tasks. Always consult this skill when the user asks to create, build, or design an n8n workflow, automate a process, or connect services — even if they don't explicitly mention 'patterns'. Covers webhook, API, database, AI, batch processing, and scheduled automation architectures.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.