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 agentmods add skills/ooooooooooooooooooop/agent-tools/evolution-proposalnpx skills add ooooooooooooooooooop/agent-tools --skill evolution-proposalgit clone --depth 1 https://github.com/ooooooooooooooooooop/agent-toolsWhat 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 | $0.00148 | $0.03429 |
| Opus 5 | $0.00074 | $0.01715 |
| Sonnet 5 | $0.00030 | $0.00686 |
| Haiku 4.5 | $0.00015 | $0.00343 |
Grade A, and why
evolution-proposal scanned grade A with 0 findings 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 yesterday.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.
进化提案(Evolution Proposal)
核心规则
本机 Agent 体系已具备进化的全部要素(可变异对象、适应度函数、选择机制、固化管线),缺的是把审计发现转化为受控变异的提案环节。本技能就是那个转化器:
- 只产提案,不直接改:输出是可评审的变更提案(diff + 证据 + 验证命令 + 预期指标),固化必须过人工或既有门禁。
- 证据驱动:每条提案必须锚定 inbox 条目中的量化证据(retries=10 / pollCount=8 / inputTokens=2.5M),无证据不立项。
- 先分析后提案:进入提案前必须先做系统化问题分析(量化基线 → 全量聚合 → 根因分类 → 优先级排序),禁止"看到一条异常就提一条提案"的碎片化处理。
- 根因分类决定对策:沿用 systematic-optimization 的三类根因——缺约束(补约束)、有约束不执行(加执行点检查/门禁)、无法强制(升级到系统层机制)。约定层不执行是经验事实,不是假设:同一反模式反复出现必须升级方案。
- 元规则隔离:AGENTS.md 治理模块(L0)是选择算子本身,本技能永远不直接改它;如需变更 L0,只能作为"人工发起的独立变更"提议,由用户执行。
适用范围与触发边界
触发(满足任一即启用本技能):
evolution-inbox(~/.agent-broker/topics/skills/evolution-inbox/workspace/inbox.jsonl)存在status: "new"的条目;- 会话中出现已知反模式信号(无 wait 轮询、重试簇、token 热点、压缩风暴)并需要根治而非临时规避;
- 用户明确要求"把最近的经验固化成规则/技能/插件改进"。
不适用:
- 单轮小修改:直接改目标文件即可,不要套提案流程(经 task-mode-router 判级);
- 无需固化的临时问答;
- L0 治理规则修改:只能由用户人工发起,本技能最多输出"建议变更内容"供用户决定。
工作流程
阶段 1:问题分析(Problem Analysis,完整九步)
本阶段完整执行 systematic-optimization 第 0~5 步(量化→全量列问题→根因→方案→联网借鉴→归纳取舍)。没有数字的"问题"是感觉;不借鉴同类已知解法的方案是闭门造车。禁止跳过本阶段直接选条目提提案。
步骤 1.1 量化基线(第 0 步)
- 读取 inbox 全部条目:
Get-Content ~/.agent-broker/topics/skills/evolution-inbox/workspace/inbox.jsonl,按行解析 JSON。 - 统计全量基线:总条目数、按 pattern 聚合的频次与占比(如
poll 26/390 = 6.7%)、按 severity 分布、按 status 分布。 - 对每个高频 pattern,统计时间分布(策略生效前/后,参考
~/.dsh/AGENTS.md各模块生效时刻),判断是历史存量还是近期增量。 - 记录基线数字(阶段 4 提案、固化后度量对比同口径使用)。
步骤 1.2 全量列问题(第 1 步)
- 按 pattern 分组,每组标注:会话数、量化证据(次数/占比/极端值)、最严重代表条目(sessionId + 具体证据)。
- 区分表象与真问题:表象是症状(如"这个会话轮询 305 次"),真问题是"缺什么机制导致这种症状反复出现"(如"监督等待阶段没有长轮询的硬约束")。
步骤 1.3 根因分类(第 2 步) 对每个 pattern 判定根因类型(沿用 systematic-optimization 三类):
| 根因类型 | 特征 | 对策 |
|---|---|---|
| 缺约束 | 根本没有对应的规则/流程/检查 | 补约束(新增规则/技能/门禁) |
| 有约束不执行 | 规则存在但执行者没遵守 | 加执行点检查(skill 硬约束、检测脚本、门禁) |
| 无法强制 | 规则无法被强制执行 | 升级到系统层机制(插件/参数/自动检测) |
判定方法:问三个问题——约束存在吗?存在但没执行吗?为什么没执行(不知道/忘了/没法强制)?第三类"无法强制"是复发问题的常见真根因:规则写在哪不重要,规则拦不拦得住才重要。
步骤 1.4 寻找解决方案(第 3 步,结构性优先) 每提出一个方案先问:这是临时方案还是结构性方案?
- 临时方案:手动清理、这次注意、下次记得(会复发);
- 结构性方案:自动检测、硬校验、参数门禁、系统拦截(无法绕过);
- 临时方案只用于止血,必须伴随结构性方案,否则问题必然复发。
What ships with it
2 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.
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.
- yesterday First seen · 152 lines · 148 tokens per session scan A f8ee9459a998
evolution-proposal is a skill published in the GitHub repository ooooooooooooooooooop/agent-tools (2 stars, last pushed 3d ago), licensed MIT. It adds 148 tokens to every session and 3,429 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
zlog
Multi-agent session log compressor and storage optimizer. Compresses .log, .out, .txt, .trace (>10KB) to .zst, .xz, or .gz (saving 77%-99.9% disk space) across /.gemini, /.config/Cursor, /.ollama, /.claude, /.windsurf, /.codex, etc. Auto-discovers AI log dirs, purges empty 0-byte logs. Safe for concurrent…
amazon-reviews-api-skill
This skill helps users automatically extract Amazon product reviews via the Amazon Reviews API. Agent should proactively apply this skill when users express needs like getting reviews for Amazon product with ASIN B07TS6R1SF, analyzing customer feedback for a specific Amazon item, getting ratings and comments for a…
agent-memory
../../../engineering/agent-memory/skills/agent-memory/SKILL.md.
ai-security
../../../engineering-team/skills/ai-security/SKILL.md.
amazon-competitor-analyzer
Scrapes Amazon product data from ASINs using browseract.com automation API and performs surgical competitive analysis. Compares specifications, pricing, review quality, and visual strategies to identify competitor moats and vulnerabilities.
ctf-malware
Provides malware analysis and network traffic techniques for CTF challenges. Use when analyzing obfuscated scripts, malicious packages, custom crypto protocols, C2 traffic, PE/.NET binaries, RC4/AES encrypted communications, YARA rules, shellcode analysis, memory forensics for malware (Volatility malfind, process…