TalentDeltaAgent templates
Every published template, in one list. Filter by sector or capability, or search by meaning.
33 templates published
Building-design agent team: an architect-lead directs 12 AEC specialists — drawing producer and reviewer, BIM/IFC automation, quantity audit, life-safety, building-performance and site-code review, construction admin — over 30 skills. A brief becomes concept drawings, layered DXF sheets, schedules and revision-impact reports. Runs in Claude Code, Codex or Gemini CLI; Python 3 drives the drawing engine. Conceptual output, not statutory approval. MIT plus a GPL-2.0-or-later Blender worker.
Agent-powered vulnerability scanner for whole repositories, not just pull requests: a fast pattern scan flags candidate files, then AI agents investigate each candidate, rate severity and record findings that can be revalidated, triaged and exported. Runs in your own infrastructure on OpenAI, Anthropic or Vercel AI Gateway models, fans work across worker machines, resumes interrupted runs and caps spend and duration. Ships a /deepsec agent skill.
Personal assistant that builds memory from what it sees and hears: six memory agents (core, episodic, semantic, procedural, resource, knowledge vault) sit under a meta-agent, with reflexion and background consolidation passes. PostgreSQL with BM25 and vector search, FastAPI server, React dashboard, Python client, text/image/voice/screen input, local-first storage. Docker Compose to start; OpenAI, Anthropic or Gemini models.
Turns a company name into a structured, source-backed briefing. A FastAPI backend runs a LangGraph pipeline — four researchers (company, financial, industry, news), then collector, curator, enricher, grounding, briefing and editor — over Tavily search, while a React interface streams progress and exports the report as PDF. Works with OpenRouter, OpenAI or Gemini; MongoDB persistence optional.
Open computer-use agent: give it a task in plain English and it looks at the screen, then clicks, types and scrolls through real desktop and web apps on macOS, Windows or Linux. Ships the S1, S2, S2.5, S3 and S4 agent generations with grounding models, behaviour best-of-N, per-agent memory, OSWorld and WindowsAgentArena harnesses, and Simular integrations.
Self-hosted YouTube channel operator: seven agents research topics, write scripts, design thumbnails, optimise SEO, render real MP4s from narration and visuals, schedule approved uploads, then fold retention and comment evidence into the next cycle. Human approval gates every publish. Node 18.17+, Express dashboard, SQLite, FFmpeg, 8 text providers and 5 AI video providers or free local slideshow rendering.
Firm-style multi-agent trading desk: four analysts (market, fundamentals, news, sentiment) run in parallel, bull and bear researchers debate them, a trader drafts the plan, three risk debaters argue it and a portfolio manager approves or rejects. Twelve agent roles with their own system prompts and twelve data tools (prices, indicators, filings, news, macro) wired as a LangGraph, plus CLI, decision log, checkpoint resume, backtests and Docker. Python 3.11+, research only: it places no orders.
An orchestrated development team of 33 Claude Code subagents: a team-configuration entry layer, core specialists (code reviewer, code archaeologist, performance optimiser), orchestrators, and full-stack, frontend, backend, mobile, DevOps, QA, security and data specialists, each with role, deliverables and hand-off contracts. Includes the CLAUDE.md team wiring, guides on creating agents and their dependency map, and the MIT LICENSE.
175 Codex CLI subagents written as TOML definitions across 13 categories, from core development, language specialists and infrastructure to quality and security, data and AI, developer experience, business and product, meta-orchestration, AI governance, platform engineering and LLMOps. Each definition carries a name, description, model guidance and a system prompt. Ships the entire repository at the pinned commit, including README, CONTRIBUTING and the MIT LICENSE.
Claude Code agent harness: a 47-plugin marketplace plus a project-local .claude/ tree — 175 agent definition files (153 distinct role names), 338 SKILL.md skills and 231 slash commands for swarm coordination, memory, research, security and cost tracking, with hooks, plugin manifests, a cross-harness .agents/ layout and an AGENTS.md entry point. Definition trees ship at the pinned commit; the v3/ next-generation source, the npm bundle and the docs site are excluded.
161 Claude Code subagents in 10 categories — core development, language specialists, infrastructure, quality and security, data and AI, developer experience and more — each a full definition with name, description, tools, model and mission body, plus 4 tool-usage agents. Ships a plugin marketplace manifest, per-category plugin manifests and an installer script.
Turns a single Claude Code session into a full game development studio: 49 agents covering game design, programming, art, audio, narrative, QA and production, plus 74 skills, 12 hooks, 13 rules and a testing framework. Scenario-driven studio workflows, department docs and a project scaffolding layer for engine work.
Multi-agent orchestration for Claude Code: 19 specialist agents, 47 skills and 21 commands that plan, decompose and run work in parallel, installed as a plugin from its own marketplace. Ships the TypeScript source and compiled dist, hooks, templates, benchmark suites and bridge tooling for Codex/OMX, with autopilot modes, task routing and verification passes.
GitHub's community collection for Copilot: 222 agent definitions under agents/, 442 skills, 194 instruction files, 8 chat workflows, 29 hooks and 104 plugin manifests, with an official plugin marketplace for one-command install. Covers language, framework, DevOps, security, data and documentation work, and ships the collection's cookbook, schemas and companion extensions.
Production-ready AI coding plugin: 68 specialised agents (language reviewers, build-error resolvers, architecture, network and security specialists), 293 workflow skills, 94 slash commands, 122 always-on rules and hook automations, mirrored across 17 harness layouts (Claude Code, Codex CLI, Cursor, Gemini, Kimi, Kiro, Qwen, Zed, Trae, OpenCode, Pi and more). Ships an MCP server catalogue defining 34 servers, Node utilities, tests and translated docs.
The complete wshobson/agents marketplace (MIT, pin 46891e7e60da): 92 plugins holding 202 coding-agent definitions, 184 skills and 105 commands - architecture, backend, frontend, APIs, databases, testing, code review, refactoring, debugging, security, CI/CD, containers, observability and docs. Ships each harness's own plugin manifest, so it installs natively on Claude Code, OpenAI Codex CLI and Cursor. 40.3k stars upstream.
A 15-agent brand-awareness outreach team curated from The Agency (github.com/msitarzewski/agency-agents, MIT, pin f99f6aa910a4): upstream's multi-channel marketing campaign runbook roster plus its Email Marketing Strategist. Social Media Strategist leads; Twitter/X, TikTok, Instagram and Reddit specialists execute; email covers segmentation, lifecycle sequences and deliverability. Brand Guardian, Trend Researcher, Analytics and Compliance agents support; the week-by-week runbook ships.
OpenClaw/ClawPod marketing agent pack adapted from Anthropic's Knowledge Work Plugins (marketing v1.2.0): SOUL identity, AGENTS workflow and 8 skills (social content, email sequences, campaign planning, brand-voice review, competitor briefs, SEO audits, performance reports) plus a TOOLS map declaring MCP connectors for email marketing (Klaviyo/Mailchimp/Brevo) and marketing automation (HubSpot/Gmail). MIT by ClawPod with Apache-2.0 notices for the derived components; pin 60c78255d16c.
Six-agent Hermes fleet for a small search and placement practice, on the Genesis Obsidian vault protocol: source material filed into schema-validated records, mailbox review, pipeline and task board work, and invoicing with a journal handoff to the accounting package. Includes the vault setup itself — 12 fileClass definitions, 9 controlled vocabularies, 9 saved views, 5 canvas dashboards, 2 guard scripts — plus 8 playbook skills. Five roles on cadences; the assistant runs on request only.
Sitebase — a Hermes agent for database-backed website development and management, for content creators. It models content in a relational database, writes the data layer its pages read, and runs the human loop: drafts with preview links, one round of review, publication only on the creator's go-ahead. Ships a SOUL, a config template and four skills, including content schema design, website data layer and ARP v1.2.1 reporting. Runtime: Hermes Agent; Hindsight memory (MIT, optional).
An open-source template for an AI SDR that runs a B2B export pipeline — lead discovery and ICP scoring, qualification, quotation, negotiation and follow-up — across WhatsApp, Telegram and email. Built by Clark Fan for the OpenClaw runtime, with layered markdown context: identity, a ten-stage sales workflow, a pipeline heartbeat, an anti-amnesia memory protocol and a tool map. Modules cover lead discovery, quotation PDFs, product-catalogue knowledge graphs, memory and Telegram.
A Hermes Agent profile: the Graphic Designer role from Kyle French's Hermes Agency pack. It composes 2D communication graphics — social cards, diagrams, posters, presentation visuals, campaign assets and layout systems — from approved brand and creative direction, and owns typography and image hierarchy, format adaptation within an approved brand system, and production exports for print and digital delivery. Brand identity strategy and campaign art direction stay with their named owners.
Open-source SEO plugin for Claude Code by Agrici Daniel. Analyst agents run in parallel across technical SEO, Core Web Vitals, Schema.org JSON-LD, E-E-A-T content quality, backlinks, local and maps intelligence and hreflang. An agentic-search track scores AI-crawler access, llms.txt and Content-Signal handling and WebMCP readiness; a GEO track covers passage-level citability in AI Overviews, ChatGPT, Perplexity and Copilot. Audits end in a prioritised action plan with falsifiability checks.
A self-evolving penetration-testing agent that runs on Kali Linux and works through 139 security skill domains, from OSINT, web and cloud exploitation to Active Directory, ICS/SCADA, RF and satellite, AI/LLM security and post-quantum crypto. Each skill carries structured payloads, test cases, a defensive perspective, detection engineering (Sigma, Splunk SPL, Sysmon, Falco) and evasion notes, with a layered memory of daily logs, distilled notes and monthly chronicles.
A daily intelligence and topic-picking system for AI and Web3 creators: it harvests signals from vendor feeds, products, markets, media, developer communities and papers, then produces a Markdown briefing with links plus a searchable single-file HTML edition. Ships 2 skills - daily-briefing and openclaw-skill-vetter - plus the identity and context set (IDENTITY, SOUL, AGENTS, USER, MEMORY, HEARTBEAT, BOOTSTRAP) as neutral templates, a Python runtime, offline sample data, docs and tests.
A Hermes profile distribution that turns a library of trading skills into an always-on research, journaling and risk-review assistant for a human US-equity trader - explicitly not an automated trading system: no order placement, no signal service, no hidden scheduled jobs, and every entry and exit decision stays with the human reviewer. Ships the profile manifest and alias, config and cron presets, ten skill bundles, schemas, English and Japanese docs, and a test suite.
A complete agent pack for the person who owns the machine: it interviews the operator on day one, wires up remote access, picks a model provider and handles keys safely, locks the box down, turns on backups and liveness monitoring, and can raise further agents — then reports to its operator in plain language. Eight skills, a provider/model config template, and a provenance record. Nothing in it is tied to any client, fleet or private deployment.
Turns a working Hermes Agent into a publishable agent pack: identity, behaviour, capabilities, tools and usage docs in one self-contained archive. It inventories a live agent, decides what ships and what stays behind, scrubs the result, classifies it against OASF and prepares a draft for review — it never publishes. Ships the set scheme, the scrub checklist, the packaging method, the ARP reporting profile, and a config template with memory optional.
Eight agent personas that work as one crew, each with its own SOUL, IDENTITY, HEARTBEAT, MEMORY, BOOTSTRAP and TOOLS files plus a shared protocol — the clearest worked example of the identity/behaviour/capability layout in this wave. Ships 34 ready skill folders, configuration templates, scripts and a setup wizard, including a path that attaches the team to an existing OpenClaw install.
A local-first code-maintenance agent whose entire core is one ReAct loop: inspect a repository, plan a change, patch it, then verify with tests, keeping an auditable record of every step. Provider-switchable over any OpenAI-compatible endpoint (DeepSeek, OpenAI, Claude or Gemini), with built-in and user-defined skill packs, MCP config examples and a 550-test suite.
A drug-discovery research agent written in Rust, with bundled skill packs for genomics, medical, pharmaceutical, research and general science work, plus MCP bridge examples, a desktop GUI and a web console. Ships install scripts for macOS, Linux and Windows and a Docker image; the README and the full docs set are in English and Chinese.
A multi-agent legal system: 67 specialist and orchestrator prompts that read documents, post findings with cited evidence, run three layers of verification and pause at a human gate before critical decisions land. Provider-agnostic (Anthropic, Mistral or local Ollama), with an audit trail, an eval suite and a one-shot setup script.
A twelve-role agent organisation modelled on a Three Departments / Six Ministries court: a triage agent sorts incoming work, a planning department drafts it, a review department can reject and send it back, and the ministries execute in parallel behind a live dashboard. Ships a SOUL file per agent, a Python backend, a web dashboard, Docker Compose and Chinese, English and Japanese setup docs.