Recently added

Captures project constraints, invariants, rules and decisions as Markdown, and injects the relevant ones back into context.

NDI's GP-07 AI Employee. Reads a card and bank transaction export, works out which charges actually recur, builds the subscription register, and flags unused seats, overlapping tools and approaching renewals — quantifying each proposed downgrade with its assumptions stated in the same row. Delivers a savings workbook. Never cancels a subscription, changes a plan or touches billing; never sums two currencies without a rate; never annualises a charge seen once; treats a security, compliance, data-retention or uncategorised tool as an escalation rather than a cancellation proposal; never presents an estimate as a realised saving; and never reports an individual person's tool usage.

An always-on YouTube brain for real estate agents — starts with a complete YouTube Game Plan (built on Mike Sherrard's frameworks), then ideation, research, SEO, scripting in the agent's own voice, lead magnets, repurposing, analytics (with an optional live-data sign-in — the only place the data connection is used), and coaching. Includes the Notion Content Dashboard (every video as a card that fills itself). Handles everything around the camera; never films, edits, or designs. Requires Plugin 1 (the Brain) set up first.

An ongoing short-form content operating system for real estate agents — daily green-screen reactions, talking-head value, and carousels, each optimized per platform (Instagram/Facebook, TikTok, YouTube Shorts). The agent thinks in formats; Claude keeps the funnel balanced, schedules posts through their own Metricool / GoHighLevel / Buffer (bring-your-own), and reads performance from live Instagram + YouTube data through ONE analytics skill — the monthly deep dive (account + competitors + gaps + a 30-day plan) plus quick reads anytime — the everyday analytics loop plus reel watch-time/skip-rate hook intelligence no scheduler exposes, plus an on-demand competitor deep dive. Text only: it writes scripts, talking points, captions, and specs — it never films, edits, or designs. Requires Plugin 1 (the Brain) set up first.

The Monthly Market System — the Market Update AI Agent. It schedules itself the moment the agent installs it, times itself to when their local real estate board actually publishes (and re-times itself as it learns), then every month pulls the fresh local numbers and builds the entire package off them: the branded PRESENTATION they screen-record (talking points under every slide — no script needed), the sendable PDF market report that doubles as a green-screen asset and the lead magnet, three short-form concepts, the Instagram 1:1 infographic AND carousel, and the distribution pack with a two-CTA newsletter. It grades last month's on-camera prediction against the numbers that just landed — the ledger segment almost no agent runs — then writes the whole month onto the same Notion Content Dashboard the YouTube and Short-Form systems use, so an unattended run is actually visible. Between runs it's a market advisor in chat — ask how any community, price range, or property type is doing, what changed since last month, what a home costs per month at today's rates, or what to tell a seller who asks if now's the time, answered instantly from the numbers already on file. Every asset ships as a paste-ready Claude Design brief with the real numbers already in it. The realtor's only job is to film. Numbers are researched once into the Brain and quoted everywhere, so nothing ever disagrees; sourced and dated or it doesn't ship, with a stated fallback for markets whose board publishes nothing. Text, data, and design briefs only — it never designs, films, or publishes. Requires Plugin 1 (the Brain) set up first.

One input — the address, or one sentence on the go ("Just listed 14 Oak — start the content") — and every listing becomes a content and lead-gen event. Eleven stages: WIN IT (the presentation + pre-listing package that gets the signature) · LIST IT (the MLS description) · LAUNCH IT (just-listed posts + Meta ad copy) · FILM IT (room-by-room YouTube walkthrough + full SEO package) · TOUR IT (listing photos → an animated cinematic home tour, no filming and no videographer) · CLIP IT (3 just-listed shorts) · SHOW IT (the open house kit) · SERVE IT (showing feedback, the weekly seller update, and the marketing report that proves the agent's value) · REVIVE IT (the honest read when it isn't selling, the price conversation, the re-launch, and the expiry conversation before the expiry) · SELL IT (the offer deadline, final calls to every agent who showed, multiple-offer process, and the under-bidder follow-up) · CLOSE IT (the just-sold content queued for closing day). Plus the brochure, the postcards, and the person-to-person outreach. Around the stages sits the system layer: a one-time setup that locks the brokerage disclaimer in so nothing ever ships a placeholder, a Listing Pipeline board in the agent's own Notion, one-command scheduling through their own Metricool / GoHighLevel / Buffer, an event-triggered morning watch (the Listing Launch AI Agent), and analytics that read back what actually worked from the agent's own closed listings. The property is captured once and never re-asked. Requires Plugin 1 (the Brain) set up first; TOUR IT needs a Higgsfield connection and hands its clips to Plugin 6 (the AI Editor).

Turns the agent's offer into a lead-capture machine. A navigator front door starts every agent the same proven way: the FIRST lead magnet is locked as the relocation guide (the #1 highest-intent search traffic in every market) with a 5-question intake pre-filled from the Brain — no overthinking; niche and brand-led guides unlock from campaign two. Then it maps the opt-in funnel that gives the guide away — the page copy section by section, a proof photo strip, and a pop-up opt-in (name, email, phone) landing on a thank-you page with the guide as an instant download. Text only: it writes the words and the strategy (a separate design skill builds the visuals); it never hosts. Lead capture only — no call booking. Requires Plugin 1 (the Brain) set up first.

The AI Editor. The agent talks to Claude in plain English; Claude directs Descript (on the agent's own account) to do the actual editing — long-form and short-form, all in one tool — while they watch it happen live. Claude is the director, not the editor. It adds the multi-source B-roll Descript can't source itself (the agent's own footage, listing photos, and free stock libraries), and a navigator skill turns vague or messy requests into one clear plan so the agent is never overwhelmed. Requires Plugin 1 (the Brain) set up first, plus a Descript connection.

One shared AI Brain for a real estate agent — their identity (brand, market, avatars, voice, offer, visuals) and their memory (clients, listings, content history) — that every skill reads from and writes back to. Run Setup once; every tool already knows you. (The AI Admin is its own plugin on this same marketplace.) Now includes the system-wide Capture layer: the agent talks on the go (a video idea, a closing, a new vendor, a market note) and it files each into the right place — ideas, proof, vendors, intel — that every other system then reads.

The Agent Leverage OS — a real estate agent's AI executive assistant, 17 skills deep. The front door understands any admin ask in plain English, then the specialists take it: Scheduling (conflict-checked booking, guests invited, video link attached, reschedules and route planning), Inbox (drafts and replies in the agent's voice), the weekday Inbox Sweep, the 7am Briefing, Client Memory (total recall of every client, note, and promise), Meeting Prep, End-of-Day Wrap, on-the-go Dispatch (just talk while mobile — it captures, acts, and reports back), Match-Back (name a property and it finds the buyers already in their notes), Chase (who owes me? — every unanswered ask ranked, with nudge drafts), Tomorrow Confirmations (the no-show killer), Filing (attachments into client folders, found in seconds), Vendor Dispatch (book the photographer in one sentence), the Showing Feedback Chaser, and the Friday Scorecard (what slipped, with the fix drafted). REQUIRES the Realtor AI Brain plugin (same marketplace): install it and run 'Set up my Brain' first — the AI Admin reads everything from the Brain and never re-interviews the agent.

PLUGIN 9 — Cohort Claude Support. The calm, always-available support desk inside Cowork. One phrase — "help" — and it takes over: answers every Claude question (Cowork vs Chat, connectors, plans and limits, permission prompts, sessions), diagnoses breakage with real look-only checks down proven decision trees (brain missing, connector expired, skill won't fire, Claude slow), teaches concepts in plain realtor English, audits setup and resumes at the first missing step, tracks the member's cohort week with a catch-up plan, and — when a human is needed — builds the complete ticket ready to paste straight into Mike's support portal. Knows all 9 plugins deeply and routes every fix to the skill that owns it; support itself is READ-ONLY by constitution. Stays current by design: money and limit answers are always fetched live from official pages, and a weekly digest cycle reads Anthropic's changelogs so answers stand on this week's Claude. Requires Plugin 1 (the Brain) for personalization; helps even without it.

A domain-agnostic methodology for building, extending, reviewing, and refactoring ontologies, domain models, taxonomies, and knowledge-graph schemas. Synthesises Stanford Ontology 101, OntoClean/UFO, Gruber's design criteria, Domain-Driven Design, OBO Foundry governance, and agent-facing platform practice into one process with explicit conflict resolution.

Sync Claude Code session data to S3-compatible storage.

Counts touched files, shows them on the statusline, speaks once at a threshold, hands the next session a machine-written handoff. Zero tokens on an ordinary turn.

Self-improving multi-agent harness that grinds your ticket backlog across every repo in your product — guarded auto-merge on green CI, and every resolved ticket writes a lesson into your git-tracked CLAUDE.md. Every ticket starts on a cheap tier and escalates once, on measured failure, instead of being predicted expensive up front; deterministic fixes skip the model entirely; and the lessons file evicts rather than accumulates, so always-on context stays a fixed budget.

Measure what an AI coding session actually costs and get advice derived from your own transcripts — where the tokens went, which were re-read on every turn, and which fixes are worth making on this machine. Tells one person to install a compressor and the next that a compressor would be worthless, because it reads their numbers first.

Policy enforcement, PII detection, and audit trails for Claude Code. Govern tool execution, scan outputs for sensitive data, and maintain compliance-grade records of every action.

Ready-to-use scientific and research Agent Skills for biology, chemistry, medicine, and related workflows.

XSEC proxy traffic workspace tools

XSEC sub-agent workspace tools

XSEC project-outcome workspace tools

XSEC project files workspace tool

XSEC conversation-tree workspace tool

XSEC browser workspace tool

XSEC session approval workspace tool

XSEC workspace system terminal

XSEC project assets, findings, reports and execution workspace

XSEC attack-path workspace and Agent tools

XSEC asset discovery business surface

Offline biological sequence/music conversion with reversible MIDI, WAV, and MusicXML, accessible score notation, and lossy arbitrary-audio transcription.

Create, illustrate, format, review, and save Chinese WeChat Official Account articles through the official draft API or authenticated browser.

사용자에게 보낼 긴 최종 답변을 독립 해석·판정하고 필요하면 다시 쓰는 Codex workflow.

사용자 표면에 나갈 최종 메시지를 무맥락 독자로 검사하고, 걸리면 다시 쓰게 한다.

Méthodologie rigoureuse de recherche en droit français, distribuée comme plugin OpenAI.

Agenrena bridge for Codex messages and experimental incoming voice calls.

Drives a Windows desktop application through UI Automation and reports what it observed: a case is a data file, an act asks the control rather than the desktop, and a check that could not run is a third verdict instead of a green.

Skill ecosystem for SS13 and BYOND Dream Maker work: a routing controller plus architecture skills for DM/SS13 coding, TGUI interfaces, modular fork content, and TGS server deployment. Engine facts are verified against the official DM Reference and live tgstation.

Deterministic task tracking for Claude Code: persists tasks the model registers via the CLI, reminds of open tasks on a message counter, and lists active tasks on session start.

Deterministic gates for Claude Code: destructive-command blocks, protected paths, delegation briefs, spec-driven flow and session-start validation. Selection lives in config, not in code.

476-tool SSH bridge for secure remote server management. Execute commands on air-gapped Linux, Windows, Docker, Kubernetes hosts via SSH from Claude Code.

crapkit's agent tools: three skills, the MCP server for score inspection and gate checks, and one advisory PostToolUse hook that names functions an edit pushed over their ceiling. The hook and the MCP server both spawn the bare `crapkit` console script, so it has to be on PATH: crapkit 0.4.0 or newer, which is the release that answers `crapkit claude-hook`. `crapkit doctor --plugin-root` compares an installed copy against the CLI it will call. The commit gate stays the only enforcement point; this hook never blocks.

Remove multi-vendor AI provenance marks: invisible Unicode (Layer A), statistical text watermarks via rewrite (Layer B), and C2PA/EXIF/XMP/container metadata on images, documents, and media. Ships the remove-ai-marks skill (thin client over the repo's HTTP service) and the self-contained clean-user-facing-text skill.

Skills de Melhoria Contínua e Excelência Operacional para o dia a dia da área. Cobre: Mapeamento e padronização · Qualidade e causa raiz · Redesenho de processo · Gestão de mudança e risco. Inclui Lean — desperdício e fluxo, Gestão da qualidade e causa raiz, Otimização de processo e Arquiteto de POP / SOP, entre outras 11.

Skills de Supply Chain e Operações para o dia a dia da área. Cobre: Planejamento de demanda e S&OP · Estoque e peças · Manutenção e frota · Roteirização e campo · Escala de equipes. Inclui Previsão de demanda, S&OP — planejamento de vendas e operações, Otimização de estoque e Planejamento de manutenção, entre outras 2.

Skills de Compras e Suprimentos para o dia a dia da área. Cobre: Estratégia de categoria · Seleção e homologação · Gestão de fornecedores · Análise de gasto. Inclui Sourcing estratégico e gestão de categoria, Análise de gasto (spend analysis), Seleção de fornecedor e Gestão de risco de fornecedor, entre outras 5.

Skills de Compliance e monitoramento regulatório para o dia a dia da área. Cobre: Integridade e anticorrupção · Monitoramento regulatório. Inclui Análise de lacuna regulatória, Monitoramento regulatório contínuo, Canal de denúncia — tratamento de relato e Monitor de regulação (alternativa).

Skills de Jurídico para o dia a dia da área. Cobre: Contratos · Contrato administrativo · Societário · Contencioso e contingência · Consultivo e triagem. Inclui Revisão de contrato contra playbook, Triagem de NDA, Revisão de NDA (avançada) e Revisão de MSA / contrato de SaaS, entre outras 18.

Skills de RH — talentos, desempenho e cultura para o dia a dia da área. Cobre: Atração e seleção · Desempenho e desenvolvimento · Estratégia de pessoas. Inclui Ciclo de avaliação de desempenho, Plano de onboarding, Descrição de vaga e Kit de entrevista estruturada, entre outras 8.

Skills de RH — DP, remuneração e relações trabalhistas para o dia a dia da área. Cobre: Departamento pessoal e legislação · Remuneração · Relações trabalhistas. Inclui Assistente CLT para RH e Departamento Pessoal, Análise de remuneração, Revisão de desligamento e Redação de política de RH, entre outras 8.

O essencial que serve a qualquer área da Tembici: conferir número antes de publicar, levar uma discussão à liderança, transformar ideia em algo que roda, e construir com guarda-corpo. Inclui Suporte a auditoria, Auditoria de planilha, Análise de gasto (spend analysis) e Escala e dimensionamento de equipes, entre outras 30.