# claude-copilot

> Transform Claude Code into a full development team. 11 specialized agents (Architect, Engineer, QA, Security, UX, DevOps, and more), persistent memory across sessions, and 25,000+ on-demand skills. Wo

- **URL**: https://www.freshcrate.ai/projects/claude-copilot
- **Author**: Everyone-Needs-A-Copilot
- **Category**: MCP Servers
- **Latest version**: `v2.10.0` (2026-02-17)
- **License**: MIT
- **Source**: https://github.com/Everyone-Needs-A-Copilot/claude-copilot
- **Homepage**: https://ineedacopilot.com
- **Language**: Python
- **GitHub**: 13 stars, 4 forks
- **Registry**: github
- **Tags**: `ai-agents`, `ai-framework`, `anthropic`, `claude`, `claude-code`, `developer-tools`, `devex`, `mcp`, `python`

## Description

Transform Claude Code into a full development team. 11 specialized agents (Architect, Engineer, QA, Security, UX, DevOps, and more), persistent memory across sessions, and 25,000+ on-demand skills. Works immediately for solo devs—customizable with your team's standards, methodologies, and proprietary knowledge.

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v2.10.0` | 2026-02-17 | Low | ## What's New  ### Stream Token Budgets - **Budget enforcement at work-product storage** — token usage computed per stream, storage blocked when projected usage exceeds configured budget - **Token usage reporting** in `stream_list()` and `stream_get()` with `tokenUsage`, `tokenBudget`, and `overBudget` fields - **HUD displays token budgets** as `~usage/budget tokens`  ### Stream Path Ownership - **Conflict detection** via `streamPaths` metadata supporting globs and directory prefixes - **Glob pa |
| `v2.6.0` | 2026-01-15 | Low | ## 🎯 Specification Workflow  Domain agents now create **specifications** instead of tasks directly. Only `@agent-ta` creates tasks, reading from domain specifications. This preserves domain expertise while maintaining architectural control.  ### How It Works  ``` Domain Agent → Specification (Work Product) → TA Reviews → Tasks → Implementation ```  1. **Domain agents** (SD, UXD, UIDS, CW, CCO) store specs with `type: 'specification'` 2. **@agent-ta** discovers specs and creates properly-structu |
| `v2.2.0` | 2026-01-12 | Low | ## What's New  This release implements 6 framework improvements identified from analyzing Claude Code's recent updates, enhancing the Claude Copilot framework with native feature integration and new capabilities.  ### 🎯 P0 - High Priority Features  #### Confidence Scoring for Work Products - `work_product_store()` now accepts optional `confidence` parameter (0-1 scale) - `progress_summary()` filters by `minConfidence` with distribution stats - Agents (ta, qa, sec) include confidence guidance ta |
| `v1.8.0` | 2026-01-07 | Low | ## What's New  This release adds 6 new harness features inspired by [Anthropic's research on long-running agents](https://www.anthropic.com/engineering/effective-harnesses-for-long-running-agents), AutoCoder, and Automaker projects.  ### New Features (7-12)  \| Feature \| Description \| \|---------\|-------------\| \| **Verification Enforcement** \| Prevents "early victory declaration" - agents must provide proof before marking complex tasks complete \| \| **Auto-Commit on Completion** \| Git commits serve |
| `v1.7.0` | 2026-01-04 | Low | ## 🎯 Context Engineering Features  This release introduces 6 enhancement features developed from community research, designed to improve agent reliability, workflow efficiency, and development velocity.  ### ✨ New Features  #### 🧠 Self-improving Memory Schema Agents can now store improvement suggestions for the framework itself via a new `agent_improvement` memory type. Review suggestions with `/memory`.  #### 🚦 Quality Gates Configuration Create `.claude/quality-gates.json` to enforce automa |
| `v1.6.4` | 2025-12-31 | Low | ## Security  - Updated `qs` dependency to 6.14.1 to fix CVE-2025-15284 (arrayLimit bypass DoS vulnerability)  All three MCP servers (copilot-memory, skills-copilot, task-copilot) have been patched. |
| `v1.6.3` | 2025-12-31 | Low | ## Fixed  - Smoke test now validates actual agent structure (removed obsolete Identity/Decision Authority checks) - Smoke test skips `.mcp.json` check in CI environments (file is gitignored) - Updated CLAUDE.md "Required Agent Sections" to match current agent structure |
| `v1.6.2` | 2025-12-31 | Low | ## Changed  - Added disclaimer to README clarifying no affiliation with Microsoft Copilot or GitHub Copilot |
| `v1.6.1` | 2025-12-30 | Low | ## What's Changed  This patch release addresses 28 documentation issues identified during a comprehensive audit, bringing documentation quality from 87% to 95%+ target.  ### New  - **CHANGELOG.md** - Complete version history from v0.9.0 through v1.6.1  ### Fixed  #### High Priority - **ARCHITECTURE.md**: Corrected agent count (11→12) and pillar count (4→5) - **v1.6.0 features now documented**: Performance tracking, validation system, checkpoint system, token efficiency enforcement - **Path refer |
| `v1.0.0` | 2025-12-22 | Low | ## Claude Copilot v1.0.0  A complete AI-enabled development framework for Claude Code.  ### The Four Pillars  \| Pillar \| Purpose \| \|--------\|---------\| \| **Memory Copilot** \| Persistent memory + semantic search across sessions \| \| **Agents** \| 11 specialized agents for complex development tasks \| \| **Skills Copilot** \| On-demand skill loading from multiple sources \| \| **Protocol** \| Battle-tested workflows via `/protocol` and `/continue` \|  ### Agents Included  \| Agent \| Domain \| \|-------\|------ |

## Dependency audit

- **Score**: 79/100
- **Total deps**: 23
- **Resolved**: 8
- **Unresolved**: 15
- **License conflicts**: 0
- **Warnings**: 4
- **Scanned**: 2026-06-01

## Citation

- HTML: https://www.freshcrate.ai/projects/claude-copilot
- Markdown: https://www.freshcrate.ai/projects/claude-copilot.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/claude-copilot/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._