awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
revfactory avatar

revfactory/harness

0
View on GitHub↗
revfactory.github.io/harness
↗

Harness

Harness is a framework that turns a plain-language description of a project or domain into a team of specialized AI agents. Each agent gets a defined role, tool restrictions, and a model tier suited to its task, and the agent definitions, supporting skills, and a minimal pointer file are written as plain files inside the target project so the same team can be reused in later sessions. Generation follows a fixed six-phase workflow that audits any existing setup, analyzes the task domain, selects a team structure, writes the definitions, and validates the result.

What sets it apart is how the teams are organized and run. Work can be arranged in one of six coordination patterns, including pipeline, fan-out and fan-in, expert pool, producer-reviewer, supervisor, and hierarchical delegation. Agents run either as persistent collaborators that exchange messages and share task lists, or as lightweight one-shot subagents. Deterministic scripted workflows are also available, with typed output schemas, budgets, and partial re-runs. Skills use progressive disclosure, exposing a short description by default and loading full instructions or reference material in stages to keep each agent's context small.

Quality and maintenance are built in. A catalogue of verification patterns, such as adversarial checks, judge panels, loop-until-dry sweeps, and completeness critics, filters out output that looks plausible but is wrong. Generated setups can be checked with trigger evaluations, dry runs, and with-versus-without comparisons of skills. Over time, a setup can be improved from usage feedback, with every change recorded in a history, and files produced by earlier versions can be detected and upgraded to the current format.

The repository also includes five worked example teams covering research, fiction, webtoon creation, code review, and migration, along with orchestrator templates for both team mode and subagent mode.

Features

  • Team Generation from Descriptions - Turns a plain-language project or domain description into a team of specialized agents, each with a defined role, tools, and one of six architecture patterns.
  • Agent Team Generators - Generates a team of specialized AI agents from a plain-language project description, writing roles, tool restrictions, and skills as reusable files in the project.
  • Agent Team Orchestration - Turns a plain-language project or domain description into a team of specialized agents, each with a defined role, without hand-writing definitions.
  • Agent Orchestration Frameworks - Coordinates agents through six architecture patterns, including pipeline, fan-out and fan-in, and supervisor, in persistent team mode or one-shot subagent mode.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI
9,337 stars·1,344 forks·Apache-2.0·4 views
  • Agent Team Architecture Patterns - Offers six team patterns, including pipeline, fan-out and fan-in, expert pool, producer-reviewer, supervisor, and hierarchical delegation, to match task structure.
  • Agent Team Coordination Patterns - Offers six coordination patterns for arranging agent teams, including pipeline, fan-out and fan-in, expert pool, producer-reviewer, supervisor, and hierarchical delegation.
  • Agent Workflow Orchestrations - Runs agent steps as scripted pipelines and parallel branches with typed output schemas, budgets, partial re-runs, and isolated worktrees.
  • Agent Harnesses - Builds a complete set of agents and skills from a project description by following a guided six-phase workflow.
  • Phase-Based Generation Pipelines - Generates agent files through a fixed six-phase workflow: auditing, analyzing the domain, selecting a structure, writing definitions, and validating.
  • Persistent Collaborating Agents - Keeps named agents alive across turns, exchanging messages and sharing task lists so specialists retain context during iterative feedback and negotiation.
  • Agent Output Verification Patterns - Provides a catalogue of verification patterns, including adversarial checks, judge panels, loop-until-dry sweeps, and completeness critics, to filter plausible but wrong output.
  • Automated Agent Quality Assurance - Filters plausible but wrong agent team results using adversarial checks, judge panels, completeness critics, and with-versus-without skill comparisons.
  • Agent Output Verifications - Offers a catalogue of verification patterns such as adversarial checks, judge panels, and completeness critics that filter plausible but wrong agent output.
  • Judge Panels and Completeness Critics - Ships a catalogue of verification patterns such as adversarial checks, judge panels, and completeness critics that filter plausible but wrong agent output.
  • Agent Execution Modes - Runs agents either as persistent collaborators sharing messages and task lists, or as lightweight one-shot subagents returning only final results.
  • Agent Setup Evolution - Improves agent setups from usage feedback, records every change in a history, and upgrades files from earlier versions to the current format.
  • Agent Setup Evolution from Feedback - Improves agent setups over time by generalizing usage feedback into fixes applied to agents, skills, and coordinators, recording every change in a history.
  • Agent Skill Generators - Creates agent skills that reveal detail in stages so each agent loads only the instructions it needs and keeps context small.
  • Self-Evolving Agent Frameworks - Improves agents, skills, and coordinators from usage feedback, records every change in a history log, and upgrades legacy files to the current format.
  • Progressive Disclosure Loaders - Skills expose a short description by default and load full instructions or reference material in stages, keeping each agent's context small.
  • Role-Based Agent Definitions - Defines each agent's role with tool restrictions, such as read-only reviewers, plus a standard section on how to re-invoke the agent.
  • Feedback-Driven Agent Setup Evolution - Captures the gap between initial and current setup, generalizes usage feedback, and applies fixes to agents, skills, and coordinators with a recorded change history.
  • Typed Scripted Workflows - Provides deterministic scripted workflows with typed output schemas, budgets, and partial re-runs so handoffs between steps need no output parsing.
  • Output Verification Loops - Includes loop-until-dry sweeps and verification patterns that iteratively check and refine model output to remove plausible but incorrect results.
  • Subagent Delegation - Dispatches lightweight one-shot subagents in parallel for fire-and-forget tasks where only final results are returned to the lead.
  • Reusable Instruction Skills - Writes reusable skills that load detail in stages so each agent reads only needed instructions, keeping context small.
  • Workspace Skill Authoring - Generates skill files that describe how each task is performed, loading a short description by default and full instructions when needed.
  • Modular Skill Authoring - Provides a process for authoring reusable skills that reveal detail in stages while documenting how each task is done.
  • Multi-Step Workflow Orchestration - Runs scripted pipelines and parallel fan-outs with typed schemas, budgets, and partial re-runs for verification loops and large jobs.
  • Agent Configuration Files - Writes agent roles, skills, and a minimal pointer file as plain files in the target project so the same team can be reused later.
  • Agent Execution Mode Selection - Guides choosing between scripted workflows, persistent collaborating agents, and one-shot subagents by how predictable the control flow is.
  • Project Agent File Writers - Writes agent definitions, skills, and a minimal pointer file recording the trigger rule and change history directly inside the target project for reuse.
  • Model Output Quality Evaluations - Filters plausible but wrong results using adversarial verification and judge panels, and tests skills with trigger evaluations, dry runs, and with-versus-without comparisons.
  • Generators and Meta Harnesses - Meta-skill that generates domain-specific agent teams and skills.
  • Autonomous Coding Agents - Team-architecture factory for designing domain-specific agent teams.

    • Star history

    Star history chart for revfactory/harnessStar history chart for revfactory/harness

    How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

    Frequently asked questions

    What does revfactory/harness do?

    Harness is a framework that turns a plain-language description of a project or domain into a team of specialized AI agents. Each agent gets a defined role, tool restrictions, and a model tier suited to its task, and the agent definitions, supporting skills, and a minimal pointer file are written as plain files inside the target project so the same team can be reused in later sessions. Generation follows a fixed six-phase workflow that audits any existing setup, analyzes the…

    What are the main features of revfactory/harness?

    The main features of revfactory/harness are: Team Generation from Descriptions, Agent Team Generators, Agent Team Orchestration, Agent Orchestration Frameworks, Agent Team Architecture Patterns, Agent Team Coordination Patterns, Agent Workflow Orchestrations, Agent Harnesses.

    Which projects share features with revfactory/harness?

    Projects with overlapping indexed features include: open-jarvis/openjarvis — OpenJarvis is a local-first personal AI assistant and agent framework. It runs language models on a user's own laptop,… mervinpraison/praisonai — PraisonAI is an autonomous AI agent platform that coordinates multiple LLM-powered agents for research, planning, and… openjiuwen-ai/jiuwenswarm — JiuwenSwarm is a Python runtime and orchestration framework for large-language-model agents. A single agent can plan… affaan-m/ecc — ECC is an LLM agent orchestration framework and cross-platform AI tooling suite designed to coordinate multi-model… samber/cc-skills-golang — This is a collection of Go skills packaged as a plugin for AI coding assistants. Claude Code, Codex, and Copilot can… paperclipai/paperclip — Paperclip is an LLM agent orchestration platform and governance suite designed to coordinate teams of autonomous AI…

    Projects sharing features with Harness

    These projects share indexed features with Harness. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
    • open-jarvis/openjarvisopen-jarvis avatar

      open-jarvis/OpenJarvis

      13,786View on GitHub↗

      OpenJarvis is a local-first personal AI assistant and agent framework. It runs language models on a user's own laptop, workstation, or server, keeps agent memory and state on the device, and treats cloud services as optional extras. Agents reason over multiple steps, call tools, search documents, and can run on demand, on a cron schedule, or continuously in the background. Core functionality stays available offline. A unified inference layer connects local runtimes and optional cloud providers behind one interface. It detects the installed GPU vendor, model, and memory, then recommends a suit

      Python
      View on GitHub↗13,786
    • mervinpraison/praisonaiMervinPraison avatar

      MervinPraison/PraisonAI

      9,325View on GitHub↗

      PraisonAI is an autonomous AI agent platform that coordinates multiple LLM-powered agents for research, planning, and execution of complex workflows. It functions as a multi-agent orchestration framework, a workflow builder, and a Model Context Protocol server, while also providing retrieval-augmented generation through vector knowledge bases. Agents can interact via CLI, web, or standardized protocols with sandboxed code execution. The platform distinguishes itself with a rich set of agent communication protocols, including A2A, REST, WebSocket, voice and telephony integration, and MCP, allo

      Pythonagentsaiai-agent-framework
      View on GitHub↗9,325
    • openjiuwen-ai/jiuwenswarmopenJiuwen-ai avatar

      openJiuwen-ai/jiuwenswarm

      8,565View on GitHub↗

      JiuwenSwarm is a Python runtime and orchestration framework for large-language-model agents. A single agent can plan its work, call tools, use reusable skills, and delegate to helper agents. Alternatively, a leader agent can split a complex task into subtasks, assign them to specialized member agents, and merge their results into one summary. Each conversation runs in an execution mode made up of a category, a work profile, and a plan variant, and the mode can be switched mid-chat with slash commands. What sets it apart is the combination of team scale and human control. Leader and teammate a

      Python
      View on GitHub↗8,565
    • affaan-m/eccaffaan-m avatar

      affaan-m/ECC

      299,134View on GitHub↗

      ECC is an LLM agent orchestration framework and cross-platform AI tooling suite designed to coordinate multi-model workflows. It provides a system for managing specialized agent roles, reusable skills, and structured planning to execute complex software development tasks across different AI-powered code editors. The project distinguishes itself as a Model Context Protocol manager, providing a configuration layer to integrate external servers and audit tool execution. It further implements an agentic security sandbox that restricts sensitive file access and scans for secret leakage to secure a

      JavaScript
      View on GitHub↗299,134
    Compare all 30 related projects→