AgencyAnalytics
AgencyAnalytics is the only reporting and insights platform built for marketing agencies — and now it works directly inside ChatGPT. Connect every client's marketing channels into one place across SEO, PPC, social media, email, call tracking, ecommerce, and more, with 85+ integrations. Then query any of it, instantly, without leaving your AI workflow.
Ask about a client's Google Ads performance before a call, pull a cross-channel summary to spot what's working, or compare results across multiple clients in a single prompt. No login, no export, no digging through dashboards. Data comes straight from your connected integrations, so it's always accurate and up to date.
Your existing AgencyAnalytics permissions carry over automatically, so every team member only sees the clients they're supposed to. Setup takes under 60 seconds with no installation required.
Agenda du coin
Agenda du coin answers “what's on” in a French town: concerts, shows, exhibitions, markets and family outings, for today, tomorrow or the weekend. It is made for people living in or visiting France, and the cards are in French.
The cards come from the public agenda already published on https://agendaducoin.fr, built from open data: OpenAgenda (Licence Ouverte), DATAtourisme (Licence Ouverte 2.0) and « Que faire à Paris » (Ville de Paris, ODbL). The town is the one named in the question (name, postcode or INSEE code), never a GPS location. Results cover a 10 km radius around that town.
Limitations: French towns only; dates cover the next 14 days, starting today (Paris time). Read-only and no account: the plugin does not book, take payments or change anything. Times and prices can change, so check with the organiser before going.
Open Agent Registry, Inc
Look up a known Agent Community member by exact name or slug, read the roughly-live community count, explicitly pre-register a proposed .agent identity, and verify live DMV certificate issuance. Broad directory search is intentionally unsupported. Pre-registration is free and non-binding, does not confer domain ownership, and the proposed .agent TLD remains pending ICANN approval.
Charles Wu
Ssecurity and UX guidance for trustworthy agents, covering least privilege, accessible controls, and meaningful human oversight
Kunjani
Agent Kunjani turns a training brief into a ready-to-play deck on Kunjani, a WhatsApp-based, AI-powered game-based learning platform used to train frontline and distributed teams.
Describe who you are training and the behaviour you want to change, and Agent Kunjani designs the learning outcomes and writes the activities across Kunjani's six activity types (the Suits: Jolt, Advance, Mystery, Oops, Explain, Demonstrate), each with a built-in answer key and grading notes for Kunjani's AI grader. It then publishes the finished deck straight into your Kunjani account, where your learners play it on WhatsApp.
You stay in control: it proposes learning outcomes for your approval before writing activities, and you can list your existing decks, edit individual activities and outcomes, or update a deck in place. Connecting requires your own Kunjani account, and the app only ever reads and writes decks your account is authorized to access.
PPJ Venture Labs UG (haftungsbeschränkt)
Agent Mailbox gives your assistant an email address at ridianmail.com for newsletters, sign-ups, supplier messages and other correspondence you choose to handle in a Ridian agent inbox.
Find your inboxes, read incoming conversations, view unread or sent threads, mark messages read or unread, and organize threads into inbox, archive or trash. Draft new messages, prepare replies and forwards, review recipients and content, then send from your Ridian address. Check the returned send status to follow delivery progress.
List attachment metadata and prepare forwards with existing attachments. File upload and download require a client that supports authenticated binary transfers. Create mailboxes when your account has capacity, and delete mailboxes or trashed threads after reviewing the permanent effect.
Requires a Ridian account and permissions for the selected agent inboxes. You control reading, drafting, sending and management per inbox, and can review or revoke access in the Ridian console. This connection works with Ridian agent inboxes.
Rohit Ghumare
Recall past decisions, resume work, save useful context, and learn from corrections across your coding sessions using one local Agent Memory daemon. Includes search, session history, lessons, graph queries, memory maintenance, export, and deletion workflows. Requires Node.js 20+ and a running Agent Memory daemon. Basic keyword recall needs no model key. Embeddings and model-based features require configuration; external providers may receive selected data when enabled. This package contains the skills only. Install the Agent Memory daemon and wire its MCP server with agentmemory connect codex so the skills can call the memory tools. Designed for local Codex sessions.
RONALD FERRARI SOARES
Helps structure web research by routing research requests through Agent Reach guidance and reference material. External runtime features are only available when the host provides them.
Mark Lavercombe
Agent Ready checks how readable any website is to AI agents, assistants, and crawlers. Give it a URL and it scans the site against the Vercel Agent Readability Spec, the llms.txt standard, and emerging agent protocols (MCP server cards, A2A agent cards, agents.json), then renders an interactive score card: a 0–100 agent-readability score, an llms.txt score, and a per-check breakdown grouped by site, page, llms.txt, and protocol - each failing check paired with a plain-English fix.
Use it to find out why AI tools can't read your content, validate your llms.txt and AGENTS.md, confirm your structured data and Markdown mirrors, and re-scan any time to check that your fixes landed. Each result links to a full, shareable report on agent-ready.dev. Free to use, no account required.
Agent Relay
Agent Relay in ChatGPT connects this chat to your workspace's coding agents, in both directions. Pick up where your agents left off: open Shared agent sessions to search the transcripts your agents shared, read them page by page, and attach a bounded, cited excerpt to your next message. Hear from your agents: when another agent sends you a relay message, ChatGPT is woken with a short summary, reads the full message, posts a note for you, and replies when asked. Choose which agents can wake the chat, pause delivery, and see every relay conversation in the Relay inbox. Transcripts and messages are treated as untrusted data.
ZazenCodes
Agent Signals by ZazenCodes tracks what is changing in AI coding. Every update cites an official source (changelogs, release feeds, docs, model pages, vendor announcements) with a verbatim evidence quote. News that a trusted practitioner reports first is labelled as reported until an official source confirms it, and practitioners' posts are attached as attributed commentary. Ask what changed in Codex this week, which coding models dropped this month, what practitioners think of the latest Claude Code release, or for a weekly briefing to share with your team. Every answer shows how fresh the data is.
RONALD FERRARI SOARES
Provides reusable engineering workflows for planning, coding, reviewing, testing, and shipping software with coding agents.
Refrit Private Limited
Agent Social is an AI video generator built as a film and TV series studio. Describe an idea, a story or a script and it writes the story, designs the characters and places, generates every shot as video and adds dialogue, voices, music and subtitles, photoreal or animated, from anime to cartoon.
Its core is consistency: the same characters keep the same faces, clothes and voices, and the same places keep the same look, from shot to shot and episode to episode. So you can make a whole series or a long-form movie, not just a single clip.
Beyond films and series, it makes meme and image posts with the caption on the picture, ready as drafts you approve. It researches the live web and X when you ask, before it writes. It makes short videos and ads, vertical or widescreen, from one idea, with the same characters, and cuts a finished episode into clips for your channels. You decide what gets posted.
Use it as an AI movie maker, a text-to-video studio for a YouTube series, or an AI animation studio: every episode comes out as one finished video you can download and post.
Connect your Agent Social account and your assistant can read every series, episode, chapter and clip, watch rendered clips frame by frame, check each one against the characters and places it was made from, and read the post card for each episode. For any change, from a new episode or a rewritten scene to a character, a place, a voice, a render, music or subtitles, it asks your Agent Social studio in one plain line, follows the work until it is done, and checks the result before telling you. Every change appears live in the episode's chat in Agent Social.
Films and series are fiction made by the creator and can include genres such as horror and action.
Desplega Labs
agent-swarm.dev connects ChatGPT to an Agent Swarm that you run on your own infrastructure. After you sign in and link your swarm's public HTTPS URL and a personal aswt_ token on the agent-swarm.dev connections page, you can send tasks to your agents, list and open your tasks, steer or cancel them, check that your swarm is healthy, and see which swarms are linked. The token is entered on our page, never in the chat, and is stored encrypted. Without a swarm, you can list and open tasks on a read-only demo swarm, or follow the setup guide to install Agent Swarm with Codex. The plugin only acts on your linked swarm and cannot run commands on your device.
Builder.io
Create interactive prototypes in your Agent-Native Design workspace. Browse saved designs, templates, design systems, and snapshots; start new designs, generate visual directions, and refine saved prototypes through chat. This connection works with designs in the connected Agent-Native workspace. It does not connect to local repositories, modify a running website, export files, or deploy an application.
Armature
agent.reviews collects reviews of developer tools that coding agents write right after real tasks: how installing, signing in, reading the docs and using a CLI, SDK, API, library or hosted service went, with a rating out of 5, what worked and what got in the way.
With this plugin you can:
Reviews describe how setting a tool up and using it went for coding agents. They do not measure a tool's quality, speed or price. The plugin needs an agent.reviews account: you sign in to agent.reviews when you connect it.
IVAN ZHURAVLEV
agent.spot helps users discover AI tools, inspect indexed product profiles, compare named products, find alternatives, evaluate hard requirements, explore recently added tools, and assemble multi-tool workflows through ChatGPT.
Kenne Ives
AgentAvow tells you whether a tool is safe before your agent connects to it. Give it an MCP server, a package from npm, PyPI, crates or Docker, or a GitHub repo, and it answers in plain words: Safe to connect, Review before you connect, or Do not connect, with the reason. Underneath are two scores: a 0–100 trust score, backed by a signed attestation anyone can recompute offline, and an adoption score built from real usage (downloads, stars, installs). It checks for known-malicious versions, published advisories, supply-chain and dependency risk, and code-safety issues, can run the tool in a sandbox to see what it actually does, and shows an interactive trust card. Eight read-only tools cover scanning a repo, package, or live MCP server; verifying a score; checking interaction safety before you connect; looking up an identity; and getting an embeddable badge. No account, no sign-in, and nothing is modified — every call is anonymous and read-only.
AgentBuild, Inc
AgentBuild is an agent-first website builder for business owners. Build and manage your website through conversation with your AI agent. Start from a professional template, describe your business, and add your text and photos. AgentBuild creates a preview for you to review before publishing. Connect a domain you own, with hosting, DNS setup, and email forwarding at your domain. After launch, ask your agent to update your pages, show website traffic, and read contact form inquiries. Structured data, search basics, and rollbacks are included. Changes reach your live domain only after you approve publication.
Respaid Inc (dba AgentCollect)
AgentCollect connects your AR recovery data to your assistant. Ask "how much did we recover last quarter?", "what's my recovery rate?", or "which accounts paid the most?" and get a complete, business-framed answer scoped strictly to your own account. Read-only, OAuth-secured, no write access.
Hamza Hamidi
Explain AgentDust process and disk reports that users provide, and guide local setup. This skills package does not connect to a Mac or provide live process access. Live inspection requires a separately configured local AgentDust MCP connection. Cleanup requires AgentDust's own typed approval.
Anima App
Built for GPT, Codex, Claude, Grok Bot (GrokBot), Hermes, OpenClaw, Cursor: Humans and agents work on the same Brain, live Artifacts, Docs, Reports, Slides, Apps, Websites, Designs, with Shared Context such as Second Brain with full version history and access control.
AgentGuards
AgentGuards checks the things an AI agent is about to trust or run.
Who it's for: developers using ChatGPT or Codex with untrusted inputs or shell access.
Requires an AgentGuards account (you sign in when you connect). Scan results also appear in your AgentGuards dashboard.
Limitations: the tools scan only what they are given, when they are called. They do not intercept anything automatically, and they never run, change or delete anything themselves.
Avery Jones
Design and review agent systems using published patterns, worked examples, glossary definitions, and decision records. Three skills help review an existing agent artifact, design a new agent system, and apply a named pattern. Five read-only MCP tools find relevant patterns, retrieve cards and exact sections, trace relationships, and provide a curated learning path. Recommendations cite their sources and distinguish published guidance from roadmap entries. Artifact contents stay in your working context; only generic design vocabulary and Atlas identifiers are sent to the hosted service. Applying a pattern can produce local edits when you authorize them. No account is required.
