[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74102":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":9,"language":10,"languages":9,"totalLinesOfCode":9,"stars":11,"forks":12,"watchers":13,"openIssues":14,"contributorsCount":15,"subscribersCount":15,"size":15,"stars1d":16,"stars7d":17,"stars30d":18,"stars90d":15,"forks30d":15,"starsTrendScore":19,"compositeScore":20,"rankGlobal":9,"rankLanguage":9,"license":9,"archived":21,"fork":21,"defaultBranch":22,"hasWiki":23,"hasPages":21,"topics":24,"createdAt":9,"pushedAt":9,"updatedAt":25,"readmeContent":26,"aiSummary":27,"trendingCount":15,"starSnapshotCount":15,"syncStatus":28,"lastSyncTime":29,"discoverSource":30},74102,"hermes-agent-self-evolution","NousResearch\u002Fhermes-agent-self-evolution","NousResearch","⚒ Evolutionary self-improvement for Hermes Agent — optimize skills, prompts, and code using DSPy + GEPA",null,"Python",4009,454,30,19,0,106,260,953,318,106.97,false,"main",true,[],"2026-06-12 04:01:13","# 🧬 Hermes Agent Self-Evolution\n\n**Evolutionary self-improvement for [Hermes Agent](https:\u002F\u002Fgithub.com\u002FNousResearch\u002Fhermes-agent).**\n\nHermes Agent Self-Evolution uses DSPy + GEPA (Genetic-Pareto Prompt Evolution) to automatically evolve and optimize Hermes Agent's skills, tool descriptions, system prompts, and code — producing measurably better versions through reflective evolutionary search.\n\n**No GPU training required.** Everything operates via API calls — mutating text, evaluating results, and selecting the best variants. ~$2-10 per optimization run.\n\n## How It Works\n\n```\nRead current skill\u002Fprompt\u002Ftool ──► Generate eval dataset\n                                        │\n                                        ▼\n                                   GEPA Optimizer ◄── Execution traces\n                                        │                    ▲\n                                        ▼                    │\n                                   Candidate variants ──► Evaluate\n                                        │\n                                   Constraint gates (tests, size limits, benchmarks)\n                                        │\n                                        ▼\n                                   Best variant ──► PR against hermes-agent\n```\n\nGEPA reads execution traces to understand *why* things fail (not just that they failed), then proposes targeted improvements. ICLR 2026 Oral, MIT licensed.\n\n## Quick Start\n\n```bash\n# Install\ngit clone https:\u002F\u002Fgithub.com\u002FNousResearch\u002Fhermes-agent-self-evolution.git\ncd hermes-agent-self-evolution\npip install -e \".[dev]\"\n\n# Point at your hermes-agent repo\nexport HERMES_AGENT_REPO=~\u002F.hermes\u002Fhermes-agent\n\n# Evolve a skill (synthetic eval data)\npython -m evolution.skills.evolve_skill \\\n    --skill github-code-review \\\n    --iterations 10 \\\n    --eval-source synthetic\n\n# Or use real session history from Claude Code, Copilot, and Hermes\npython -m evolution.skills.evolve_skill \\\n    --skill github-code-review \\\n    --iterations 10 \\\n    --eval-source sessiondb\n```\n\n## What It Optimizes\n\n| Phase | Target | Engine | Status |\n|-------|--------|--------|--------|\n| **Phase 1** | Skill files (SKILL.md) | DSPy + GEPA | ✅ Implemented |\n| **Phase 2** | Tool descriptions | DSPy + GEPA | 🔲 Planned |\n| **Phase 3** | System prompt sections | DSPy + GEPA | 🔲 Planned |\n| **Phase 4** | Tool implementation code | Darwinian Evolver | 🔲 Planned |\n| **Phase 5** | Continuous improvement loop | Automated pipeline | 🔲 Planned |\n\n## Engines\n\n| Engine | What It Does | License |\n|--------|-------------|---------|\n| **[DSPy](https:\u002F\u002Fgithub.com\u002Fstanfordnlp\u002Fdspy) + [GEPA](https:\u002F\u002Fgithub.com\u002Fgepa-ai\u002Fgepa)** | Reflective prompt evolution — reads execution traces, proposes targeted mutations | MIT |\n| **[Darwinian Evolver](https:\u002F\u002Fgithub.com\u002Fimbue-ai\u002Fdarwinian_evolver)** | Code evolution with Git-based organisms | AGPL v3 (external CLI only) |\n\n## Guardrails\n\nEvery evolved variant must pass:\n1. **Full test suite** — `pytest tests\u002F -q` must pass 100%\n2. **Size limits** — Skills ≤15KB, tool descriptions ≤500 chars\n3. **Caching compatibility** — No mid-conversation changes\n4. **Semantic preservation** — Must not drift from original purpose\n5. **PR review** — All changes go through human review, never direct commit\n\n## Full Plan\n\nSee [PLAN.md](PLAN.md) for the complete architecture, evaluation data strategy, constraints, benchmarks integration, and phased timeline.\n\n## License\n\nMIT — © 2026 Nous Research\n","Hermes Agent Self-Evolution 是一个用于优化 Hermes Agent 技能、提示和代码的自动进化项目。该项目利用 DSPy 和 GEPA（遗传帕累托提示进化）技术，通过反射性进化搜索来不断改进 Hermes Agent 的各项功能，无需 GPU 训练，所有操作均通过 API 调用来完成，包括文本变异、结果评估和最优变体选择，单次优化成本约为 2-10 美元。适用于需要持续优化和改进 AI 代理技能及其相关代码的场景，如代码审查等任务。目前，项目已实现对技能文件的优化，并计划逐步扩展到工具描述、系统提示和代码实现等多个方面。",2,"2026-06-11 03:48:49","high_star"]