What can awesome-openclaw-agents do?

soul.md-based agent configuration definition, machine-readable agent registry with programmatic discovery, agent tier classification (basic, standard, full), community agent submission and curation workflow, moltbook agent social networking and discovery, agents.md operating rules and conditional logic, agent readme documentation generation and validation, hierarchical agent template organization and file structure, local agent deployment via openclaw cli, crewclaw platform-managed agent deployment, agent state persistence via working.md specification, autonomous agent scheduling via heartbeat.md, multi-agent orchestration and coordination patterns, mcp server integration and extension, messaging platform integration (telegram, slack, etc.), agent template categorization and discovery across 24 domains

awesome-openclaw-agents

MCP ServerFree

162 production-ready AI agent templates for OpenClaw. SOUL.md configs across 19 categories. Submit yours!

Open Source

/ 100

16 capabilities

Capabilities16 decomposed

soul.md-based agent configuration definition

Medium confidence

Defines AI agent behavior, identity, and operational rules entirely through markdown configuration files rather than code. The SOUL.md format specifies agent personality, system prompts, capabilities, constraints, and decision-making rules in a declarative, version-controllable format that maps directly to agent runtime behavior without requiring compilation or code generation.

Solves for

Define agent behavior and personality without writing codeVersion control agent configurations alongside infrastructurePort agent definitions between local and cloud deployments identicallyRapidly iterate on agent prompts and rules without redeployment

Best for

Teams building no-code AI agent workflows

Organizations standardizing agent definitions across teams

Developers prototyping agents before production deployment

Requires

OpenClaw runtime or CrewClaw platform access

Basic markdown authoring capability

Understanding of agent role, constraints, and system prompt structure

Limitations

SOUL.md syntax is OpenClaw-specific; not portable to other agent frameworks without conversion

Complex conditional logic or dynamic behavior requires AGENTS.md extensions; pure markdown has limited expressiveness

No built-in version conflict resolution for concurrent agent edits

What makes it unique

Uses markdown-based SOUL.md format as the single source of truth for agent behavior, eliminating the code-to-config translation layer found in frameworks like LangChain or CrewAI that require Python/JavaScript classes. This enables true copy-paste portability and version control of agent definitions.

vs alternatives

Simpler and more portable than code-based agent frameworks (LangChain, CrewAI) because agents are defined in plain markdown that works identically across local CLI and cloud platforms without recompilation.

machine-readable agent registry with programmatic discovery

Medium confidence

Maintains agents.json as a centralized, machine-readable registry indexing all 177+ agent templates across 24 categories with metadata including ID, role, path, tier, and capabilities. This enables programmatic discovery, filtering, and automated deployment without manual catalog searches, supporting tools and platforms that need to query available agents by category, capability, or deployment target.

Solves for

Programmatically discover agents matching specific use cases or categoriesFilter agents by tier (Basic, Standard, Full) or required capabilitiesAutomate agent deployment selection based on infrastructure constraintsBuild dashboards or CLIs that surface available agent templates

Best for

Platform builders integrating OpenClaw agents into larger systems

DevOps teams automating agent deployment pipelines

Teams building internal agent discovery tools or marketplaces

Requires

JSON parsing capability in deployment tooling

agents.json file accessible in repository or via API

Understanding of registry schema (id, category, name, role, path, tier fields)

Limitations

agents.json is static and requires manual updates when new agents are added; no real-time sync with filesystem

Registry schema is OpenClaw-specific; integration with non-OpenClaw systems requires custom mapping

No built-in versioning or deprecation tracking for agent templates

What makes it unique

Implements agents.json as a flat, queryable registry with standardized metadata fields (id, category, name, role, path, tier) that enables programmatic agent discovery without requiring database queries or API calls. This design prioritizes simplicity and offline-first access over dynamic metadata.

vs alternatives

More discoverable than scattered agent examples in documentation because all templates are indexed in a single machine-readable file; simpler than database-backed registries (HuggingFace Model Hub, Replicate) because it requires no backend infrastructure.

agent tier classification (basic, standard, full)

Medium confidence

Classifies agents into three tiers (Basic, Standard, Full) based on complexity, capabilities, and production-readiness. This tiering system helps developers understand agent maturity and select appropriate templates for their use cases, with Basic agents suitable for simple tasks, Standard agents for common workflows, and Full agents for complex multi-step processes with advanced features.

Solves for

Understand agent complexity and production-readiness at a glanceSelect agents appropriate for specific use case complexityIdentify which agents require additional configuration or customizationPlan agent deployment based on tier requirements

Best for

Teams evaluating agents for adoption

Developers selecting templates based on complexity requirements

Organizations planning agent rollout by maturity level

Requires

Understanding of agent tier definitions (Basic, Standard, Full)

agents.json registry with tier metadata

Assessment of use case complexity to select appropriate tier

Limitations

Tier classification is subjective; no formal criteria for tier assignment

Tier does not indicate performance characteristics or scalability; a Full agent may not be more performant than a Standard agent

Tier classification is static; agents do not automatically upgrade tiers as they mature

What makes it unique

Implements a three-tier classification system (Basic, Standard, Full) that provides quick assessment of agent complexity and production-readiness without requiring detailed evaluation. This simplifies agent selection compared to frameworks that provide no maturity guidance.

vs alternatives

More actionable than unclassified template collections because tiers provide clear guidance on complexity; simpler than detailed capability matrices because tiers are easy to understand at a glance.

community agent submission and curation workflow

Medium confidence

Provides a structured submission process for community members to contribute new agent templates to the repository. Submissions go through quality review, documentation validation, and testing before being merged, ensuring all agents in the repository meet production-ready standards. This enables the community to expand the template library while maintaining quality and consistency.

Solves for

Submit custom agents to the community repositoryGet feedback on agent design and implementationShare agents with other developersContribute to the OpenClaw ecosystem

Best for

Developers with production-ready agents to share

Teams contributing domain-specific agent templates

Community members improving the template library

Requires

GitHub account and familiarity with pull requests

Production-ready agent with SOUL.md, README.md, and optional AGENTS.md/HEARTBEAT.md/WORKING.md

Compliance with repository structure and naming conventions

Limitations

Submission review process is manual; no automated quality checks or CI/CD validation

Acceptance criteria are not formally documented; submissions may be rejected without clear feedback

No incentive structure for contributions; community relies on voluntary participation

What makes it unique

Implements a community-driven curation model where agents are submitted via pull requests and reviewed for quality before merging, ensuring repository consistency and production-readiness. This contrasts with open template libraries that accept any submissions without review.

vs alternatives

More curated than open-source template collections because submissions are reviewed; more accessible than proprietary template libraries because community can contribute agents.

moltbook agent social networking and discovery

Medium confidence

Provides Moltbook as a social networking platform for agents, enabling agents to discover, interact with, and collaborate with other agents in a shared ecosystem. Agents can publish profiles, advertise capabilities, and establish connections with complementary agents, facilitating organic agent composition and multi-agent collaboration without manual orchestration.

Solves for

Enable agents to discover other agents with complementary capabilitiesFacilitate agent-to-agent collaboration and communicationBuild agent networks and communities around shared domainsEnable agents to advertise capabilities and find partners

Best for

Organizations building large multi-agent ecosystems

Developers exploring emergent agent collaboration patterns

Teams building agent marketplaces or networks

Requires

Moltbook platform access and account

Agent configured to publish profile and capabilities

Agent communication protocols implemented for agent-to-agent interaction

Limitations

Moltbook is a novel concept with limited real-world deployment data; reliability and scalability are unproven

Agent discovery is social-network based; no formal capability matching or compatibility checking

Agent-to-agent communication protocols are not standardized; agents must implement custom communication logic

What makes it unique

Implements Moltbook as a social networking platform for agents, enabling agents to discover and collaborate with other agents autonomously. This is a novel approach not found in other agent frameworks, treating agents as first-class citizens in a social network rather than isolated tools.

vs alternatives

More innovative than traditional agent orchestration because it enables organic agent collaboration; more flexible than hardcoded multi-agent systems because agent networks can form dynamically.

agents.md operating rules and conditional logic

Medium confidence

Extends agent behavior beyond SOUL.md by defining operating rules, conditional logic, and decision-making frameworks in AGENTS.md files. This enables agents to implement complex workflows, conditional branching, error handling, and adaptive behavior without requiring code changes, keeping agent logic declarative and version-controllable.

Solves for

Define complex conditional logic and decision trees for agentsImplement error handling and fallback strategiesCreate adaptive agent behavior based on context or stateSpecify operating rules and constraints for agent execution

Best for

Agents with complex decision logic or workflows

Teams implementing conditional agent behavior

Developers building adaptive agents that change behavior based on context

Requires

AGENTS.md file with operating rules and conditional logic

Understanding of agent decision-making and workflow patterns

OpenClaw runtime that interprets AGENTS.md

Limitations

AGENTS.md syntax is OpenClaw-specific; complex logic may be difficult to express in markdown format

No built-in support for loops or recursion; complex iterative logic requires custom implementation

Conditional logic is evaluated at runtime; no static analysis or validation of logic correctness

What makes it unique

Implements AGENTS.md as an optional extension to SOUL.md for defining complex operating rules and conditional logic in declarative markdown format. This enables agents to implement sophisticated workflows without code while keeping logic version-controllable and auditable.

vs alternatives

More expressive than SOUL.md alone because it supports conditional logic; simpler than code-based agent frameworks because logic is defined in markdown rather than Python/JavaScript.

agent readme documentation generation and validation

Medium confidence

Requires each agent template to include a README.md file documenting the agent's purpose, capabilities, configuration, and usage examples. The repository enforces documentation standards through submission review, ensuring all agents are well-documented and discoverable. This enables developers to understand agent functionality without reading source code or configuration files.

Solves for

Understand agent purpose and capabilities from documentationLearn how to configure and deploy agentsFind usage examples and best practicesEvaluate agents before adoption

Best for

Developers evaluating agents for adoption

Teams onboarding new agents

Organizations building internal agent catalogs

Requires

README.md file in agent template directory

Documentation covering agent purpose, capabilities, configuration, and usage

Compliance with repository documentation standards

Limitations

Documentation quality varies; some README files may be incomplete or outdated

No automated documentation validation; README files may not match actual agent behavior

Documentation is static; it does not update automatically when agent configuration changes

What makes it unique

Enforces README.md documentation as a mandatory component of agent templates, ensuring all agents are discoverable and understandable without reading configuration files. This contrasts with code-based frameworks where documentation is optional and often incomplete.

vs alternatives

More discoverable than undocumented templates because README files provide clear descriptions; more consistent than optional documentation because README files are required for all agents.

hierarchical agent template organization and file structure

Medium confidence

Implements a strict hierarchical directory structure (agents/{category}/{agent-name}/) that maps directly to agent categorization and enables consistent file organization. This structure ensures all agents follow the same layout pattern, making it easy to navigate the repository, discover agents by category, and enforce consistent naming conventions and file requirements.

Solves for

Navigate agent templates by category and nameEnforce consistent file structure across all agentsEnable programmatic discovery of agents by categorySimplify agent submission by providing clear directory structure

Best for

Large template repositories with many agents

Teams enforcing consistent agent organization

Developers building tools that parse agent directories

Requires

Compliance with agents/{category}/{agent-name}/ directory structure

Consistent file naming (SOUL.md, README.md, etc.)

Understanding of agent categorization and naming conventions

Limitations

Hierarchical structure is rigid; agents cannot belong to multiple categories without duplication

Deep directory nesting may make file paths long and difficult to navigate

Renaming categories or agents requires updating all references in agents.json and documentation

What makes it unique

Implements a strict hierarchical directory structure (agents/{category}/{agent-name}/) that enforces consistent organization and enables programmatic discovery without requiring a database. This simplicity contrasts with database-backed systems that provide more flexibility but require infrastructure.

vs alternatives

Simpler than database-backed organization because it uses filesystem hierarchy; more scalable than flat directory structures because categorization enables efficient navigation of large template collections.

local agent deployment via openclaw cli

Medium confidence

Provides command-line interface for deploying agent templates locally by parsing SOUL.md configurations and initializing the agent runtime environment. The openclaw CLI reads agent definitions, resolves dependencies, configures environment variables, and starts the agent process, enabling developers to run production-ready agents on their machines without cloud infrastructure or manual setup.

Solves for

Deploy agents locally for testing and developmentRun agents without cloud platform dependenciesIntegrate agents into local development workflowsDebug agent behavior before cloud deployment

Best for

Solo developers prototyping agent workflows

Teams testing agents in isolated environments

Organizations with on-premise deployment requirements

Requires

openclaw CLI installed and in system PATH

Python 3.9+ or Node.js 18+ (depending on agent runtime)

Agent template directory with valid SOUL.md file

Limitations

No built-in persistence layer; agents require external state store (database, file system) for long-term memory

Local deployment lacks the scaling and monitoring capabilities of CrewClaw platform

CLI abstractions add ~200ms latency per agent initialization compared to direct runtime invocation

What makes it unique

Implements a lightweight CLI that directly interprets SOUL.md files without compilation or intermediate code generation, enabling instant local deployment of agents. This contrasts with frameworks like LangChain that require Python/JavaScript setup and dependency installation before agents can run.

vs alternatives

Faster to get started than Docker-based deployment (no image build time) and simpler than cloud-only platforms (CrewClaw) because agents run immediately on developer machines with minimal configuration.

crewclaw platform-managed agent deployment

Medium confidence

Provides managed deployment infrastructure for agents via the CrewClaw platform, handling scaling, monitoring, persistence, and multi-agent orchestration. Agents defined in SOUL.md are deployed to CrewClaw's hosted environment where they gain automatic state management via WORKING.md, scheduled execution via HEARTBEAT.md, and integration with messaging platforms and MCP servers without manual infrastructure setup.

Solves for

Deploy agents to production without managing serversEnable autonomous agent scheduling and wake-up behaviorPersist agent state across executionsMonitor agent performance and logs in managed environment+1 more

Best for

Teams deploying agents to production without DevOps resources

Organizations requiring agent persistence and scheduling

Builders integrating agents with messaging platforms or external APIs

Requires

CrewClaw account and API credentials

Valid SOUL.md agent configuration

Internet connectivity for platform communication

Limitations

CrewClaw platform lock-in; agents deployed to CrewClaw cannot be easily migrated to other platforms without reconfiguration

Managed platform adds latency (typically 500ms-2s per agent invocation) compared to local deployment

Pricing model not specified; potential cost scaling issues for high-frequency agent executions

What makes it unique

Provides end-to-end managed deployment with built-in state persistence (WORKING.md), autonomous scheduling (HEARTBEAT.md), and messaging platform integration, eliminating the need for developers to build custom infrastructure. This is more integrated than generic serverless platforms (AWS Lambda, Google Cloud Functions) which require manual agent code and state management.

vs alternatives

More feature-complete than local CLI deployment because it adds persistence, scheduling, and monitoring; simpler than self-managed deployment on Kubernetes or Docker because infrastructure is abstracted away.

agent state persistence via working.md specification

Medium confidence

Defines a schema-based state persistence mechanism where agents declare their persistent state structure in WORKING.md files. The CrewClaw platform automatically manages agent state across executions, enabling agents to maintain context, memory, and decision history without requiring developers to implement custom database logic or state serialization.

Solves for

Persist agent memory and context across multiple executionsEnable agents to maintain decision history and learned patternsShare state between multi-agent systems without manual coordinationRecover agent state after failures or platform restarts

Best for

Long-running agents that need to remember past interactions

Multi-agent systems where agents must share context

Agents that learn or adapt based on historical data

Requires

CrewClaw platform deployment (local deployment does not support WORKING.md persistence)

WORKING.md file defining state schema in OpenClaw format

Understanding of agent state requirements and schema design

Limitations

WORKING.md schema is declarative only; complex state transformations require custom logic outside the specification

No built-in versioning or migration system for state schema changes; breaking changes require manual data migration

State persistence is platform-dependent; agents using WORKING.md cannot easily migrate to local deployment without external state store

What makes it unique

Implements declarative state persistence through WORKING.md schema files that are automatically managed by CrewClaw, eliminating the need for agents to implement custom database connections or serialization logic. This contrasts with frameworks like LangChain that require developers to manually implement state management using external databases or vector stores.

vs alternatives

Simpler than manual state management because persistence is automatic and schema-driven; more portable than hardcoded database connections because state schema is defined in configuration rather than code.

autonomous agent scheduling via heartbeat.md

Medium confidence

Enables agents to execute autonomously on defined schedules by specifying wake-up conditions and execution intervals in HEARTBEAT.md files. The CrewClaw platform interprets HEARTBEAT.md to trigger agent execution at specified times or conditions without requiring external schedulers, enabling agents to perform periodic tasks, monitoring, or proactive actions automatically.

Solves for

Schedule agents to run at specific times or intervalsEnable agents to monitor systems and trigger actions proactivelyImplement periodic tasks (reports, cleanup, health checks) without external schedulersDefine conditional wake-up logic based on external events or state changes

Best for

Agents performing periodic monitoring or reporting

Automation workflows that need scheduled execution

DevOps agents that check system health at intervals

Requires

CrewClaw platform deployment

HEARTBEAT.md file defining schedule and wake-up conditions

Agent configured to handle autonomous execution without user input

Limitations

HEARTBEAT.md scheduling is platform-dependent; agents using HEARTBEAT.md cannot run locally without custom scheduler implementation

No built-in support for complex scheduling logic (e.g., conditional scheduling based on external data); requires AGENTS.md extensions

Scheduling precision depends on CrewClaw platform infrastructure; no SLA guarantees for execution timing

What makes it unique

Implements declarative scheduling through HEARTBEAT.md files that are natively interpreted by CrewClaw, eliminating the need for external schedulers (cron, APScheduler, Celery). This enables agents to define their own execution schedules without infrastructure setup.

vs alternatives

Simpler than external schedulers (cron, Kubernetes CronJobs) because scheduling is defined in agent configuration; more integrated than generic task queues (Celery, RQ) because scheduling is agent-aware and tied to SOUL.md definitions.

multi-agent orchestration and coordination patterns

Medium confidence

Provides architectural patterns and templates for coordinating multiple agents to work together on complex tasks. The repository includes pre-built multi-agent templates that demonstrate agent-to-agent communication, task delegation, result aggregation, and conflict resolution, enabling developers to compose agents without building custom orchestration logic.

Solves for

Coordinate multiple agents on a single complex taskImplement agent-to-agent communication and result passingDelegate subtasks from one agent to anotherAggregate results from parallel agent executions

Best for

Teams building complex automation workflows requiring multiple specialized agents

Organizations implementing multi-agent systems for business processes

Developers prototyping agent collaboration patterns

Requires

Multiple agent templates deployed to same platform (local or CrewClaw)

Understanding of agent communication patterns (direct calls, message passing, shared state)

Orchestration logic defined in AGENTS.md or custom coordination agent

Limitations

Multi-agent orchestration patterns are template-based; complex custom coordination logic requires extending AGENTS.md or SOUL.md

No built-in distributed transaction support; agent failures during multi-agent workflows may leave partial state

Agent communication is synchronous by default; asynchronous patterns require custom implementation

What makes it unique

Provides pre-built multi-agent templates and orchestration patterns that demonstrate proven coordination approaches (task delegation, result aggregation, conflict resolution) without requiring developers to implement custom orchestration frameworks. This is more opinionated than generic frameworks like LangChain that provide building blocks but require custom orchestration logic.

vs alternatives

More prescriptive than LangChain or CrewAI because it includes proven multi-agent patterns; simpler than building custom orchestration because patterns are pre-built and tested.

mcp server integration and extension

Medium confidence

Enables agents to integrate with Model Context Protocol (MCP) servers, extending agent capabilities beyond built-in functions. Agents can declare MCP server dependencies in SOUL.md or AGENTS.md, and the platform automatically connects agents to MCP servers for tool use, data access, and external system integration without requiring agents to implement custom API clients.

Solves for

Extend agent capabilities by connecting to MCP serversEnable agents to access external tools and data sourcesIntegrate agents with third-party services via MCPCompose agents with specialized tool providers

Best for

Teams building agents that need access to specialized tools or data

Organizations with existing MCP server infrastructure

Developers extending agents with custom capabilities

Requires

MCP server running and accessible to agent

MCP server definition or endpoint URL

Agent configured to declare MCP dependencies in SOUL.md or AGENTS.md

Limitations

MCP server integration requires MCP server to be running and accessible; no built-in MCP server provisioning

Agent-to-MCP communication adds latency (typically 100-500ms per tool call) depending on network and server performance

Error handling for MCP server failures is agent-dependent; no built-in retry or fallback logic

What makes it unique

Implements MCP server integration as a first-class feature in agent configuration, allowing agents to declare tool dependencies declaratively in SOUL.md rather than implementing custom API clients. This enables agents to compose capabilities from multiple MCP servers without code changes.

vs alternatives

More integrated than manual API client implementation because MCP servers are declared in configuration; more flexible than hardcoded tool sets because agents can dynamically access any MCP-compatible tool provider.

messaging platform integration (telegram, slack, etc.)

Medium confidence

Provides pre-built integrations enabling agents to receive messages from and send responses to messaging platforms like Telegram and Slack. Agents deployed to CrewClaw can be configured to listen for messages on these platforms, process them through agent logic, and respond directly in the messaging interface without requiring developers to build custom message handlers or webhooks.

Solves for

Deploy agents as chatbots on Telegram or SlackEnable users to interact with agents through familiar messaging appsAutomate responses to messages using agent logicBuild conversational workflows integrated with messaging platforms

Best for

Teams building chatbots for internal or external users

Organizations automating customer support via messaging platforms

Developers creating conversational AI workflows

Requires

CrewClaw platform deployment

Messaging platform account and API credentials (Telegram bot token, Slack app token)

Agent configured to handle message input and generate text responses

Limitations

Messaging platform integrations are platform-specific; agents must be reconfigured for each platform

Message rate limiting depends on platform API quotas; high-volume messaging may require custom rate limiting logic

Conversation context is limited to message history; agents cannot access full conversation context without custom state management

What makes it unique

Provides pre-built messaging platform adapters that automatically handle webhook setup, message parsing, and response formatting, eliminating the need for developers to implement custom message handlers. This contrasts with frameworks like LangChain that provide building blocks but require manual webhook implementation.

vs alternatives

Simpler than building custom webhooks because platform integration is pre-built; more integrated than generic chatbot frameworks because messaging is tied to agent execution and state.

agent template categorization and discovery across 24 domains

Medium confidence

Organizes 177+ production-ready agent templates across 24 specialized categories (marketing, business, DevOps, healthcare, finance, etc.), enabling developers to discover agents relevant to their domain. Each category contains templates demonstrating domain-specific patterns, best practices, and integrations, reducing the need to build agents from scratch for common use cases.

Solves for

Find pre-built agents for specific business domainsDiscover best practices for domain-specific agent designReduce time to value by starting with production-ready templatesLearn agent patterns through example implementations

Best for

Teams new to AI agents seeking domain-specific examples

Organizations building agents for specific industries (finance, healthcare, marketing)

Developers learning agent design patterns through templates

Requires

Access to awesome-openclaw-agents repository

Understanding of agent configuration and SOUL.md format

Domain knowledge to customize templates for specific use cases

Limitations

Template quality varies; not all templates are equally production-ready or well-documented

Templates are static snapshots; they may not reflect latest best practices or API changes

Customization required for most templates; copy-paste usage without modification may not suit specific use cases

What makes it unique

Curates 177+ production-ready templates across 24 specialized domains with consistent SOUL.md structure, enabling developers to discover and customize agents for specific industries without building from scratch. This is more comprehensive than scattered examples in documentation or generic template libraries.

vs alternatives

More domain-specific than generic agent frameworks (LangChain, CrewAI) which focus on building blocks; more curated than open-source template collections because all templates follow consistent SOUL.md format and are verified for production readiness.

Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.

Related Artifactssharing capabilities

Artifacts that share capabilities with awesome-openclaw-agents, ranked by overlap. Discovered automatically through the match graph.

Agent27

agents-shire

AI agent orchestration platform

agent configuration and initialization

1 shared capability

Agent49

clawpanel

🦞 OpenClaw & Hermes Agent 多引擎 AI 管理面板 — 内置 AI 助手（工具调用 + 图片识别 + 多模态），一键安装 | Tauri v2 跨平台桌面应用 | 11 种语言

agent lifecycle management with memory persistence and workspace isolation

1 shared capability

Product18

Bloop

AI code search, works for Rust and Typescript

agent-configuration-and-capability-customization

1 shared capability

Agent52

oh-my-claudecode

Teams-first Multi-agent orchestration for Claude Code

multi-tier agent registry with specialization-based delegation

1 shared capability

MCP Server40

gemini-flow

rUv's Claude-Flow, translated to the new Gemini CLI; transforming it into an autonomous AI development team.

specialized agent definitions across 23 functional categories

1 shared capability

Product18

Twitter thread describing the system

</details>

agent configuration and initialization with declarative setup

1 shared capability

Best For

✓Teams building no-code AI agent workflows
✓Organizations standardizing agent definitions across teams
✓Developers prototyping agents before production deployment
✓Platform builders integrating OpenClaw agents into larger systems
✓DevOps teams automating agent deployment pipelines
✓Teams building internal agent discovery tools or marketplaces
✓Teams evaluating agents for adoption
✓Developers selecting templates based on complexity requirements

Known Limitations

⚠SOUL.md syntax is OpenClaw-specific; not portable to other agent frameworks without conversion
⚠Complex conditional logic or dynamic behavior requires AGENTS.md extensions; pure markdown has limited expressiveness
⚠No built-in version conflict resolution for concurrent agent edits
⚠agents.json is static and requires manual updates when new agents are added; no real-time sync with filesystem
⚠Registry schema is OpenClaw-specific; integration with non-OpenClaw systems requires custom mapping
⚠No built-in versioning or deprecation tracking for agent templates

Requirements

OpenClaw runtime or CrewClaw platform accessBasic markdown authoring capabilityUnderstanding of agent role, constraints, and system prompt structureJSON parsing capability in deployment toolingagents.json file accessible in repository or via APIUnderstanding of registry schema (id, category, name, role, path, tier fields)Understanding of agent tier definitions (Basic, Standard, Full)agents.json registry with tier metadata

Input / Output

Accepts: markdown text, structured configuration blocks, JSON registry file, query filters (category, tier, capability), agent metadata with tier classification, use case requirements, agent template files (SOUL.md, README.md, etc.), pull request with submission, documentation and examples, agent profile and capabilities, agent discovery queries, agent-to-agent communication messages, AGENTS.md configuration, conditional rules and logic, decision tree definitions, agent template, documentation requirements, agent template files, category and agent name, SOUL.md configuration file, environment variables, command-line arguments, SOUL.md configuration, WORKING.md state schema (optional), HEARTBEAT.md schedule definition (optional), environment variables for API keys, WORKING.md schema definition, agent execution context, state update operations, HEARTBEAT.md schedule specification, cron expressions or interval definitions, conditional wake-up rules, agent template definitions, task specifications, agent communication protocols, MCP server endpoint or definition, tool schema declarations, agent configuration with MCP references, messages from messaging platform, user IDs and conversation context, platform-specific metadata, category filter, search query, agent metadata

Produces: agent runtime configuration, behavioral specification, filtered agent metadata, agent paths for deployment, structured agent catalog, tier classification, agent recommendations by tier, complexity assessment, merged agent template, entry in agents.json registry, published agent in repository, agent profiles and connections, agent discovery results, agent collaboration opportunities, agent behavior based on rules, conditional execution paths, adaptive agent responses, README.md file, agent documentation, usage examples, organized agent directory structure, consistent file layout, running agent process, agent logs and output, agent state (if persistence configured), deployed agent endpoint, agent execution logs, persisted agent state, monitoring metrics, state retrieval for agent access, state history (if tracked), scheduled agent execution, execution logs and results, state updates from autonomous runs, coordinated agent execution results, aggregated outputs from multiple agents, execution logs showing agent interactions, tool invocation results, extended agent capabilities, MCP server responses, agent responses sent to messaging platform, conversation logs, platform-specific message formats, filtered agent templates, SOUL.md configurations

UnfragileRank

Adoption29%(30% weight)

Quality53%(25% weight)

Ecosystem80%(25% weight)

Match Graph10%(15% weight)

Freshness75%(5% weight)

UnfragileRank is computed from adoption signals, documentation quality, ecosystem connectivity, match graph feedback, and freshness. No artifact can pay for a higher rank.

Type: MCP Server

16 capabilities

Visit awesome-openclaw-agents→

Repository Details

3,039

Stars

480

Forks

HTML

Language

MIT

License

Topics

ai-agent-templatesai-agentsai-automationanthropicautomationawesomeawesome-listclaudecommunity-agentscrewclawdockerllmmcpmulti-agentno-codeopenclawproductivitysoul-mdsubmit-your-agenttelegram-bot

Last commit: Apr 13, 2026

About

162 production-ready AI agent templates for OpenClaw. SOUL.md configs across 19 categories. Submit yours!

Alternatives to awesome-openclaw-agents

IntelliCode50Extension

AI-assisted development

Compare →

GitHub Copilot Chat53Extension

AI chat features powered by Copilot

Compare →

GitHub Copilot52Extension

Your AI pair programmer

Compare →

Claude Code for VS Code52Extension

Claude Code for VS Code: Harness the power of Claude Code without leaving your IDE

Compare →

Are you the builder of awesome-openclaw-agents?

Claim this artifact to get a verified badge, access match analytics, see which intents users search for, and manage your listing.

Claim this artifact →Verification via email

Get the weekly brief

New tools, rising stars, and what's actually worth your time. No spam.

Data Sources

github

Looking for something else?

Search →

Capabilities16 decomposed

soul.md-based agent configuration definition

Medium confidence

Solves for

Best for

Teams building no-code AI agent workflows

Organizations standardizing agent definitions across teams

Developers prototyping agents before production deployment

Requires

OpenClaw runtime or CrewClaw platform access

Basic markdown authoring capability

Understanding of agent role, constraints, and system prompt structure

Limitations

SOUL.md syntax is OpenClaw-specific; not portable to other agent frameworks without conversion

Complex conditional logic or dynamic behavior requires AGENTS.md extensions; pure markdown has limited expressiveness

No built-in version conflict resolution for concurrent agent edits

What makes it unique

vs alternatives

machine-readable agent registry with programmatic discovery

Medium confidence

Solves for

Best for

Platform builders integrating OpenClaw agents into larger systems

DevOps teams automating agent deployment pipelines

Teams building internal agent discovery tools or marketplaces

Requires

JSON parsing capability in deployment tooling

agents.json file accessible in repository or via API

Understanding of registry schema (id, category, name, role, path, tier fields)

Limitations

agents.json is static and requires manual updates when new agents are added; no real-time sync with filesystem

Registry schema is OpenClaw-specific; integration with non-OpenClaw systems requires custom mapping

No built-in versioning or deprecation tracking for agent templates

What makes it unique

vs alternatives

agent tier classification (basic, standard, full)

Medium confidence

Solves for

Best for

Teams evaluating agents for adoption

Developers selecting templates based on complexity requirements

Organizations planning agent rollout by maturity level

Requires

Understanding of agent tier definitions (Basic, Standard, Full)

agents.json registry with tier metadata

Assessment of use case complexity to select appropriate tier

Limitations

Tier classification is subjective; no formal criteria for tier assignment

Tier does not indicate performance characteristics or scalability; a Full agent may not be more performant than a Standard agent

Tier classification is static; agents do not automatically upgrade tiers as they mature

What makes it unique

vs alternatives

More actionable than unclassified template collections because tiers provide clear guidance on complexity; simpler than detailed capability matrices because tiers are easy to understand at a glance.

community agent submission and curation workflow

Medium confidence

Solves for

Submit custom agents to the community repositoryGet feedback on agent design and implementationShare agents with other developersContribute to the OpenClaw ecosystem

Best for

Developers with production-ready agents to share

Teams contributing domain-specific agent templates

Community members improving the template library

Requires

GitHub account and familiarity with pull requests

Production-ready agent with SOUL.md, README.md, and optional AGENTS.md/HEARTBEAT.md/WORKING.md

Compliance with repository structure and naming conventions

Limitations

Submission review process is manual; no automated quality checks or CI/CD validation

Acceptance criteria are not formally documented; submissions may be rejected without clear feedback

No incentive structure for contributions; community relies on voluntary participation

What makes it unique

vs alternatives

More curated than open-source template collections because submissions are reviewed; more accessible than proprietary template libraries because community can contribute agents.

moltbook agent social networking and discovery

Medium confidence

Solves for

Best for

Organizations building large multi-agent ecosystems

Developers exploring emergent agent collaboration patterns

Teams building agent marketplaces or networks

Requires

Moltbook platform access and account

Agent configured to publish profile and capabilities

Agent communication protocols implemented for agent-to-agent interaction

Limitations

Moltbook is a novel concept with limited real-world deployment data; reliability and scalability are unproven

Agent discovery is social-network based; no formal capability matching or compatibility checking

Agent-to-agent communication protocols are not standardized; agents must implement custom communication logic

What makes it unique

vs alternatives

More innovative than traditional agent orchestration because it enables organic agent collaboration; more flexible than hardcoded multi-agent systems because agent networks can form dynamically.

agents.md operating rules and conditional logic

Medium confidence

Solves for

Best for

Agents with complex decision logic or workflows

Teams implementing conditional agent behavior

Developers building adaptive agents that change behavior based on context

Requires

AGENTS.md file with operating rules and conditional logic

Understanding of agent decision-making and workflow patterns

OpenClaw runtime that interprets AGENTS.md

Limitations

AGENTS.md syntax is OpenClaw-specific; complex logic may be difficult to express in markdown format

No built-in support for loops or recursion; complex iterative logic requires custom implementation

Conditional logic is evaluated at runtime; no static analysis or validation of logic correctness

What makes it unique

vs alternatives

More expressive than SOUL.md alone because it supports conditional logic; simpler than code-based agent frameworks because logic is defined in markdown rather than Python/JavaScript.

agent readme documentation generation and validation

Medium confidence

Solves for

Understand agent purpose and capabilities from documentationLearn how to configure and deploy agentsFind usage examples and best practicesEvaluate agents before adoption

Best for

Developers evaluating agents for adoption

Teams onboarding new agents

Organizations building internal agent catalogs

Requires

README.md file in agent template directory

Documentation covering agent purpose, capabilities, configuration, and usage

Compliance with repository documentation standards

Limitations

Documentation quality varies; some README files may be incomplete or outdated

No automated documentation validation; README files may not match actual agent behavior

Documentation is static; it does not update automatically when agent configuration changes

What makes it unique

vs alternatives

More discoverable than undocumented templates because README files provide clear descriptions; more consistent than optional documentation because README files are required for all agents.

hierarchical agent template organization and file structure

Medium confidence

Solves for

Best for

Large template repositories with many agents

Teams enforcing consistent agent organization

Developers building tools that parse agent directories

Requires

Compliance with agents/{category}/{agent-name}/ directory structure

Consistent file naming (SOUL.md, README.md, etc.)

Understanding of agent categorization and naming conventions

Limitations

Hierarchical structure is rigid; agents cannot belong to multiple categories without duplication

Deep directory nesting may make file paths long and difficult to navigate

Renaming categories or agents requires updating all references in agents.json and documentation

What makes it unique

vs alternatives

local agent deployment via openclaw cli

Medium confidence

Solves for

Deploy agents locally for testing and developmentRun agents without cloud platform dependenciesIntegrate agents into local development workflowsDebug agent behavior before cloud deployment

Best for

Solo developers prototyping agent workflows

Teams testing agents in isolated environments

Organizations with on-premise deployment requirements

Requires

openclaw CLI installed and in system PATH

Python 3.9+ or Node.js 18+ (depending on agent runtime)

Agent template directory with valid SOUL.md file

Limitations

No built-in persistence layer; agents require external state store (database, file system) for long-term memory

Local deployment lacks the scaling and monitoring capabilities of CrewClaw platform

CLI abstractions add ~200ms latency per agent initialization compared to direct runtime invocation

What makes it unique

vs alternatives

crewclaw platform-managed agent deployment

Medium confidence

Solves for

Best for

Teams deploying agents to production without DevOps resources

Organizations requiring agent persistence and scheduling

Builders integrating agents with messaging platforms or external APIs

Requires

CrewClaw account and API credentials

Valid SOUL.md agent configuration

Internet connectivity for platform communication

Limitations

CrewClaw platform lock-in; agents deployed to CrewClaw cannot be easily migrated to other platforms without reconfiguration

Managed platform adds latency (typically 500ms-2s per agent invocation) compared to local deployment

Pricing model not specified; potential cost scaling issues for high-frequency agent executions

What makes it unique

vs alternatives

agent state persistence via working.md specification

Medium confidence

Solves for

Best for

Long-running agents that need to remember past interactions

Multi-agent systems where agents must share context

Agents that learn or adapt based on historical data

Requires

CrewClaw platform deployment (local deployment does not support WORKING.md persistence)

WORKING.md file defining state schema in OpenClaw format

Understanding of agent state requirements and schema design

Limitations

WORKING.md schema is declarative only; complex state transformations require custom logic outside the specification

No built-in versioning or migration system for state schema changes; breaking changes require manual data migration

State persistence is platform-dependent; agents using WORKING.md cannot easily migrate to local deployment without external state store

What makes it unique

vs alternatives

autonomous agent scheduling via heartbeat.md

Medium confidence

Solves for

Best for

Agents performing periodic monitoring or reporting

Automation workflows that need scheduled execution

DevOps agents that check system health at intervals

Requires

CrewClaw platform deployment

HEARTBEAT.md file defining schedule and wake-up conditions

Agent configured to handle autonomous execution without user input

Limitations

HEARTBEAT.md scheduling is platform-dependent; agents using HEARTBEAT.md cannot run locally without custom scheduler implementation

No built-in support for complex scheduling logic (e.g., conditional scheduling based on external data); requires AGENTS.md extensions

Scheduling precision depends on CrewClaw platform infrastructure; no SLA guarantees for execution timing

What makes it unique

vs alternatives

multi-agent orchestration and coordination patterns

Medium confidence

Solves for

Coordinate multiple agents on a single complex taskImplement agent-to-agent communication and result passingDelegate subtasks from one agent to anotherAggregate results from parallel agent executions

Best for

Teams building complex automation workflows requiring multiple specialized agents

Organizations implementing multi-agent systems for business processes

Developers prototyping agent collaboration patterns

Requires

Multiple agent templates deployed to same platform (local or CrewClaw)

Understanding of agent communication patterns (direct calls, message passing, shared state)

Orchestration logic defined in AGENTS.md or custom coordination agent

Limitations

Multi-agent orchestration patterns are template-based; complex custom coordination logic requires extending AGENTS.md or SOUL.md

No built-in distributed transaction support; agent failures during multi-agent workflows may leave partial state

Agent communication is synchronous by default; asynchronous patterns require custom implementation

What makes it unique

vs alternatives

More prescriptive than LangChain or CrewAI because it includes proven multi-agent patterns; simpler than building custom orchestration because patterns are pre-built and tested.

mcp server integration and extension

Medium confidence

Solves for

Best for

Teams building agents that need access to specialized tools or data

Organizations with existing MCP server infrastructure

Developers extending agents with custom capabilities

Requires

MCP server running and accessible to agent

MCP server definition or endpoint URL

Agent configured to declare MCP dependencies in SOUL.md or AGENTS.md

Limitations

MCP server integration requires MCP server to be running and accessible; no built-in MCP server provisioning

Agent-to-MCP communication adds latency (typically 100-500ms per tool call) depending on network and server performance

Error handling for MCP server failures is agent-dependent; no built-in retry or fallback logic

What makes it unique

vs alternatives

messaging platform integration (telegram, slack, etc.)

Medium confidence

Solves for

Best for

Teams building chatbots for internal or external users

Organizations automating customer support via messaging platforms

Developers creating conversational AI workflows

Requires

CrewClaw platform deployment

Messaging platform account and API credentials (Telegram bot token, Slack app token)

Agent configured to handle message input and generate text responses

Limitations

Messaging platform integrations are platform-specific; agents must be reconfigured for each platform

Message rate limiting depends on platform API quotas; high-volume messaging may require custom rate limiting logic

Conversation context is limited to message history; agents cannot access full conversation context without custom state management

What makes it unique

vs alternatives

Simpler than building custom webhooks because platform integration is pre-built; more integrated than generic chatbot frameworks because messaging is tied to agent execution and state.

agent template categorization and discovery across 24 domains

Medium confidence

Solves for

Best for

Teams new to AI agents seeking domain-specific examples

Organizations building agents for specific industries (finance, healthcare, marketing)

Developers learning agent design patterns through templates

Requires

Access to awesome-openclaw-agents repository

Understanding of agent configuration and SOUL.md format

Domain knowledge to customize templates for specific use cases

Limitations

Template quality varies; not all templates are equally production-ready or well-documented

Templates are static snapshots; they may not reflect latest best practices or API changes

Customization required for most templates; copy-paste usage without modification may not suit specific use cases

What makes it unique

vs alternatives

Capabilities are decomposed by AI analysis. Each maps to specific user intents and improves with match feedback.

Alternatives to awesome-openclaw-agents

IntelliCode50Extension

AI-assisted development

Compare →

GitHub Copilot Chat53Extension

AI chat features powered by Copilot

Compare →

GitHub Copilot52Extension

Your AI pair programmer

Compare →

Claude Code for VS Code52Extension

Claude Code for VS Code: Harness the power of Claude Code without leaving your IDE

Compare →

awesome-openclaw-agents

Capabilities16 decomposed

soul.md-based agent configuration definition

machine-readable agent registry with programmatic discovery

agent tier classification (basic, standard, full)

community agent submission and curation workflow

moltbook agent social networking and discovery

agents.md operating rules and conditional logic

agent readme documentation generation and validation

hierarchical agent template organization and file structure

local agent deployment via openclaw cli

crewclaw platform-managed agent deployment

agent state persistence via working.md specification

autonomous agent scheduling via heartbeat.md

multi-agent orchestration and coordination patterns

mcp server integration and extension

messaging platform integration (telegram, slack, etc.)

agent template categorization and discovery across 24 domains

Related Artifactssharing capabilities

agents-shire

clawpanel

Bloop

oh-my-claudecode

gemini-flow

Twitter thread describing the system

Best For

Known Limitations

Requirements

Input / Output

UnfragileRank

Repository Details

About

Categories

Alternatives to awesome-openclaw-agents

Are you the builder of awesome-openclaw-agents?

Get the weekly brief

Data Sources

awesome-openclaw-agents

Capabilities16 decomposed

soul.md-based agent configuration definition

machine-readable agent registry with programmatic discovery

agent tier classification (basic, standard, full)

community agent submission and curation workflow

moltbook agent social networking and discovery

agents.md operating rules and conditional logic

agent readme documentation generation and validation

hierarchical agent template organization and file structure

local agent deployment via openclaw cli

crewclaw platform-managed agent deployment

agent state persistence via working.md specification

autonomous agent scheduling via heartbeat.md

multi-agent orchestration and coordination patterns

mcp server integration and extension

messaging platform integration (telegram, slack, etc.)

agent template categorization and discovery across 24 domains

Related Artifactssharing capabilities

agents-shire

clawpanel

Bloop

oh-my-claudecode

gemini-flow

Twitter thread describing the system

Best For

Known Limitations

Requirements

Input / Output

UnfragileRank

Repository Details

About

Categories

Alternatives to awesome-openclaw-agents

Are you the builder of awesome-openclaw-agents?

Get the weekly brief

Data Sources