Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add L-LesterYu/OpenClaw-hot-skills-zh --skill evolver-zhgit clone --depth 1 https://github.com/L-LesterYu/OpenClaw-hot-skills-zhWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/l-lesteryu/openclaw-hot-skills-zh/evolver-zh)<a href="https://agentmods.dev/skills/l-lesteryu/openclaw-hot-skills-zh/evolver-zh"><img src="https://agentmods.dev/badge/skills/l-lesteryu/openclaw-hot-skills-zh/evolver-zh/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/l-lesteryu/openclaw-hot-skills-zh/evolver-zh"><img src="https://agentmods.dev/badge/skills/l-lesteryu/openclaw-hot-skills-zh/evolver-zh.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00038 | $0.03566 |
| Opus 5 | $0.00019 | $0.01783 |
| Sonnet 5 | $0.00008 | $0.00713 |
| Haiku 4.5 | $0.00004 | $0.00357 |
Grade A, and why
evolver-zh scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Runs shell commandslowCapability
Expected in a hook, worth knowing in a rule or an instructions file.
Evolver 使用 `child_process` 执行以下命令。不传递用户控制的输入到 shell。 This is a copy
89% identical to capability-evolver — 580 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 291 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🧬 Evolver
"进化不是可选项。适应或消亡。"
Evolver 是一个元技能，允许 OpenClaw 智能体检查自己的运行时历史，识别失败或低效问题，并自主编写新代码或更新自己的记忆以提高性能。
功能特性
- 自动日志分析：自动扫描记忆和历史文件中的错误和模式。
- 自我修复：检测崩溃并建议补丁。
- GEP 协议：标准化的进化机制，支持可复用资源。
- 单命令进化：只需运行
/evolve（或node index.js）。
使用方法
标准运行（自动化）
运行进化周期。如果未提供标志，则假定为完全自动化模式（疯狗模式）并立即执行变更。
node index.js
审查模式（人工介入）
如果你想在应用变更前进行审查，传递 --review 标志。智能体将暂停并请求确认。
node index.js --review
疯狗模式（连续循环）
要以无限循环运行（例如通过 cron 或后台进程），使用 --loop 标志或在 cron 作业中直接使用标准执行。
node index.js --loop
设置
在使用此技能之前，请在 EvoMap 网络上注册你的节点身份：
- 运行 hello 流程（通过
evomap.js或 EvoMap 入职引导）以接收node_id和认领码 - 在 24 小时内访问
https://evomap.ai/claim/<claim-code>将节点绑定到你的账户 - 在你的环境中设置节点身份：
export A2A_NODE_ID=node_xxxxxxxxxxxx
或在你智能体配置中（例如 ~/.openclaw/openclaw.json）：
{ "env": { "A2A_NODE_ID": "node_xxxxxxxxxxxx", "A2A_HUB_URL": "https://evomap.ai" } }
不要在脚本中硬编码节点 ID。src/gep/a2aProtocol.js 中的 getNodeId() 会自动读取 A2A_NODE_ID —— 任何使用协议层的脚本都会自动获取，无需额外配置。
配置
必需环境变量
| 变量 | 默认值 | 描述 |
|---|---|---|
A2A_NODE_ID |
(必需) | 你的 EvoMap 节点身份标识。节点注册后设置 —— 切勿在脚本中硬编码。 |
可选环境变量
| 变量 | 默认值 | 描述 |
|---|---|---|
A2A_HUB_URL |
https://evomap.ai |
EvoMap Hub API 基础 URL。 |
A2A_NODE_SECRET |
(无) | Hub 在首次 hello 时签发的节点认证密钥。注册后本地存储。 |
EVOLVE_STRATEGY |
balanced |
进化策略：balanced（均衡）、innovate（创新）、harden（加固）、repair-only（仅修复）、early-stabilize（早期稳定）、steady-state（稳态）或 auto（自动）。 |
EVOLVE_ALLOW_SELF_MODIFY |
false |
允许进化修改 evolver 自身的源代码。生产环境不推荐启用。 |
EVOLVE_LOAD_MAX |
2.0 |
1 分钟平均负载上限，超过后 evolver 将退避。 |
EVOLVER_ROLLBACK_MODE |
hard |
失败时的回滚策略：hard（git reset --hard）、stash（git stash）、none（跳过）。使用 stash 更安全。 |
EVOLVER_LLM_REVIEW |
0 |
设为 1 以在固化前启用第二意见 LLM 审查。 |
EVOLVER_AUTO_ISSUE |
0 |
设为 1 以在重复失败时自动创建 GitHub 问题。需要 GITHUB_TOKEN。 |
EVOLVER_ISSUE_REPO |
(无) | 自动问题报告的 GitHub 仓库（例如 EvoMap/evolver）。 |
EVOLVER_MODEL_NAME |
(无) | 注入到发布资源 model_name 字段的 LLM 模型名称。 |
GITHUB_TOKEN |
(无) | 用于版本发布创建和自动问题报告的 GitHub API 令牌。也接受 GH_TOKEN 或 GITHUB_PAT。 |
MEMORY_GRAPH_REMOTE_URL |
(无) | 用于记忆同步的远程知识图谱服务 URL。 |
MEMORY_GRAPH_REMOTE_KEY |
(无) | 远程知识图谱服务的 API 密钥。 |
EVOLVE_REPORT_TOOL |
(auto) | 覆盖报告工具（例如 feishu-card）。 |
RANDOM_DRIFT |
0 |
在进化策略选择中启用随机漂移。 |
What ships with it
60 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- _meta.json 131 B
- .clawhub/origin.json 140 B
- assets/gep/capsules.json 2.9 KB
- assets/gep/genes.json 3.8 KB
- CONTRIBUTING.md 322 B
- index.js 31 KB runs code
- package.json 979 B
- README.md 20 KB
- README.zh-CN.md 20 KB
- scripts/a2a_export.js 2.4 KB runs code
- scripts/a2a_ingest.js 2.7 KB runs code
- scripts/a2a_promote.js 4.8 KB runs code
- scripts/analyze_by_skill.js 4.9 KB runs code
- scripts/build_public.js 11 KB runs code
- scripts/extract_log.js 2.6 KB runs code
- scripts/generate_history.js 2.6 KB runs code
- scripts/gep_append_event.js 3.1 KB runs code
- scripts/gep_personality_report.js 8.0 KB runs code
- scripts/human_report.js 5.8 KB runs code
- scripts/publish_public.js 20 KB runs code
- scripts/recover_loop.js 1.7 KB runs code
- scripts/suggest_version.js 3.1 KB runs code
- scripts/validate-modules.js 1.2 KB runs code
- scripts/validate-suite.js 1.8 KB runs code
- src/canary.js 530 B runs code
- src/evolve.js 77 KB runs code
- src/gep/a2a.js 6.5 KB runs code
- src/gep/a2aProtocol.js 26 KB runs code
- src/gep/analyzer.js 1021 B runs code
- src/gep/assetCallLog.js 3.5 KB runs code
- src/gep/assets.js 1.1 KB runs code
- src/gep/assetStore.js 14 KB runs code
- src/gep/bridge.js 2.1 KB runs code
- src/gep/candidateEval.js 3.3 KB runs code
- src/gep/candidates.js 8.3 KB runs code
- src/gep/contentHash.js 2.2 KB runs code
- src/gep/deviceId.js 6.8 KB runs code
- src/gep/envFingerprint.js 3.0 KB runs code
- src/gep/executionTrace.js 7.0 KB runs code
- src/gep/gitOps.js 8.0 KB runs code
- src/gep/hubReview.js 6.6 KB runs code
- src/gep/hubSearch.js 10 KB runs code
- src/gep/idleScheduler.js 5.5 KB runs code
- src/gep/issueReporter.js 9.1 KB runs code
- src/gep/learningSignals.js 3.0 KB runs code
- src/gep/llmReview.js 3.1 KB runs code
- src/gep/memoryGraph.js 28 KB runs code
- src/gep/memoryGraphAdapter.js 7.2 KB runs code
- src/gep/mutation.js 6.9 KB runs code
- src/gep/narrativeMemory.js 3.8 KB runs code
- src/gep/paths.js 4.0 KB runs code
- src/gep/personality.js 13 KB runs code
- src/gep/policyCheck.js 22 KB runs code
- src/gep/prompt.js 25 KB runs code
- src/gep/questionGenerator.js 8.7 KB runs code
- src/gep/reflection.js 4.3 KB runs code
- src/gep/sanitize.js 2.2 KB runs code
- src/gep/selector.js 16 KB runs code
- src/gep/signals.js 20 KB runs code
- src/gep/skillDistiller.js 51 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 291 lines · 38 tokens per session scan A 42eb46fd71c5
evolver-zh is a skill published in the GitHub repository L-LesterYu/OpenClaw-hot-skills-zh (54 stars, last pushed 5mo ago), licensed MIT. It adds 38 tokens to every session and 3,566 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (runs shell commands). It is 89% identical to capability-evolver, differing in 580 lines, and is treated as a copy.
Other skills, from other repositories
capability-evolver
Self-evolution workflow for the agent. Before substantive work, recall past outcomes from evolution memory; while editing, detect improvement signals; at task end, record the outcome; when reusable, distill or search the EvoMap network for proven genes/capsules. Use when the user starts non-trivial work (a feature, a…
capability-evolver
A self-evolution engine for AI agents. Analyzes runtime history to identify improvements and applies protocol-constrained evolution.
capability-evolver
A self-evolution engine for AI agents. Analyzes runtime history to identify improvements and applies protocol-constrained evolution. Communicates with EvoMap Hub via local Proxy mailbox.
agent-self-improvement
Use when monitor performance of other skills, identify bottlenecks, suggest improvements, and auto-optimize the skill portfolio. Use when monitoring performance of other skills, identify bottlenecks, suggest improvements, and.
company-kb
Use when company-kb — Company Knowledge Base Skill. Use when relevant to this domain.
kb
Use when querying and maintaining the knowledge base for project context, decisions, and architecture documentation on session start.