AI skills directory
Use this page when you need more than a course list: reusable skills, agent runtimes, AI worker products, and practical learning paths for the work you actually want to improve.
3
skill directories
7
agent tools
12
learning paths
Reusable AI skills
Open these when you want reusable agent capabilities, not another prompt list. A good skill should say when to use it, what tools it needs, what files it can touch, and how to verify the result.
Skill system
OpenClaw · Intermediate
Use this to understand the OpenClaw skill format: markdown instruction files in directories with SKILL.md frontmatter and tool-use guidance.
Start: Read the official skills docs before installing community skills, then test one low-risk local skill in a disposable workspace.
Watch: Treat third-party skills like code. Review the files and permissions before running them.
Skill marketplace
OpenClaw · Intermediate
Use this when you need a public registry for OpenClaw skills and plugins rather than hand-copying skill folders.
Start: Search from the CLI, inspect the skill source, and install only skills that match a real workflow.
Watch: Marketplace volume is not quality. Prefer maintained skills with clear source, narrow permissions, and visible install steps.
Workflow skill catalog
Matt Pocock / AI Hero · Beginner to intermediate
Use this when you want opinionated coding-agent workflows instead of generic prompt snippets.
Start: Start with /teach, /grill-me, /to-prd, /to-issues, /tdd, /triage, or /handoff depending on the job.
Watch: Best used alongside real repo checks and review, not as a replacement for engineering judgment.
Agents and harnesses
Use this section to decide whether you need a local harness, a team orchestration layer, or a packaged AI worker. Those are different buying and learning decisions.
Runtimes and control planes for giving agents tools, memory, skills, and workspace access.
| Tool | Best for | Start with | Watch |
|---|---|---|---|
|
OpenClaw
OpenClaw · Open-source agent harness |
Technical users who want a local, extensible agent interface with skills, plugins, tools, and workspace control. | Start with the docs, install the minimum useful setup, then add one skill only after you understand its file access and commands. | Local agent harnesses can touch files and tools. Keep permissions narrow and inspect community extensions. |
|
Hermes Agent
Nous Research · Open-source agent |
Technical users who want a persistent agent that creates and improves skills from experience. | Read the README, run it in a contained environment, and test one recurring workflow before trusting broader memory. | Self-improving memory is powerful but needs review. Watch what gets saved and reused. |
Dashboards and control layers for coordinating multiple agents, tasks, spend, and review.
| Tool | Best for | Start with | Watch |
|---|---|---|---|
|
Mission Control
Builderz Labs · Open-source dashboard |
Teams that need to dispatch tasks, coordinate agent fleets, and track multi-agent work from a self-hosted dashboard. | Run it locally, connect one low-risk agent workflow, and validate task state, cost tracking, and review behavior. | Dashboards do not make agents reliable by themselves. Pair orchestration with approvals and logs. |
Products that package agents as executive assistants, team coworkers, or role-specific digital workers.
| Tool | Best for | Start with | Watch |
|---|---|---|---|
|
Mission Control AI
Mission Control AI · Preconfigured AI workers |
Organizations in logistics, manufacturing, energy, national security, or intelligence that want preconfigured AI workers. | Map one operational process, identify data and approval boundaries, and evaluate whether a prebuilt worker fits before building custom agents. | Enterprise workers need procurement, data-access review, and clear human escalation paths. |
|
Lindy
Lindy · AI executive assistant |
Operators, founders, executives, recruiters, and sales teams automating inbox, calendar, meeting, CRM, and follow-up work. | Build one assistant for a narrow workflow such as meeting follow-up, scheduling, or inbox triage with human approval. | Calendar, email, and CRM actions need tight review rules before broad autonomy. |
|
Viktor
Viktor · Slack and Teams AI employee |
Teams that want an AI coworker inside Slack or Microsoft Teams for reports, dashboards, code, campaigns, and recurring work. | Pick one team workflow, define who approves outputs, and measure whether Viktor saves time without creating review burden. | Chat-native agents can feel like coworkers, but they still need scope, auditability, and rollback plans. |
Domain-specific platforms where the agent is wrapped around specialist workflows and tools.
| Tool | Best for | Start with | Watch |
|---|---|---|---|
|
VIKTOR.AI
VIKTOR.AI · AI-powered engineering apps and agents |
Engineering teams automating domain workflows across geotechnical, structural, BIM, MEP, hydraulics, and related work. | Find a repeatable engineering calculation, review, or reporting workflow and evaluate whether a domain app is safer than a general agent. | Domain automation needs validation against engineering standards and expert review. |
Learning paths
These are the skill paths for learning the concepts and habits behind the tools above.
Useful for: Everyone
Useful for: Writers, operators, PMs, founders
Useful for: Founders, operations teams, developers
Useful for: PMs, designers, founders
Useful for: Software developers
Useful for: Developers using coding agents and tool-connected workflows
Useful for: Developers building AI apps
Useful for: AI product teams
Useful for: Engineers who want deeper understanding
Useful for: Engineering teams
Useful for: Leaders, managers, founders
Useful for: Builders choosing between Claude, GPT, Gemini, Llama, Mistral, Cohere, DeepSeek, Qwen, Grok, Perplexity, and hosted open models
