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 TashanGKD/cognitive-os --skill cognitive-integrate-fragmentsgit clone --depth 1 https://github.com/TashanGKD/cognitive-osWrote 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/tashangkd/cognitive-os/cognitive-integrate-fragments)<a href="https://agentmods.dev/skills/tashangkd/cognitive-os/cognitive-integrate-fragments"><img src="https://agentmods.dev/badge/skills/tashangkd/cognitive-os/cognitive-integrate-fragments/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/tashangkd/cognitive-os/cognitive-integrate-fragments"><img src="https://agentmods.dev/badge/skills/tashangkd/cognitive-os/cognitive-integrate-fragments.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 | $0.00073 | $0.03716 |
| Opus 5 | $0.00036 | $0.01858 |
| Sonnet 5 | $0.00015 | $0.00743 |
| Haiku 4.5 | $0.00007 | $0.00372 |
Grade A, and why
cognitive-integrate-fragments 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 11d 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.
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.
This is a copy
88% identical to cognitive-integrate-fragments — 10 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 — 260 lines — stays where its author put it; the contents beside it link to each section on GitHub.
碎片整合 Skill(Integrate Fragments)
实现「小人整合机制」:读取待整合碎片 → 判断同化/顺应 → 生成具体更新建议 → 用户确认 → 执行写入 → 级联更新。
知识导航表(执行前必须理解的概念根)
| 层级 | 文档 | 需要理解的概念 |
|---|---|---|
| D0 认知根(必读) | cognitive/L1_knowledge/formal_spec/self_evolving_agent_spec.md |
层5:ceremony(K) = 备份+修改+版本+级联;层3:C2关系显式化(修改K1后必须通知依赖它的B-objects) |
| D3 规范参考 | cognitive/maintenance_protocol.md |
K1文档修改规范:版本管理/历史备份/矛盾检测要求 |
| D4 运行时数据 | cognitive/L2_fragments/fragment_index.md + 目标L1文档 |
待整合碎片清单 + 目标文档当前内容(整合前必须理解已有结构) |
核心概念速查: ① 小人机制 = 以「已在L1文档里居住的知识」视角判断碎片如何融入,不破坏已有结构 ② 整合 = K-object升级:fragment→被L1吸收,碎片整合索引更新为🔲→✅ ③ 整合后必须更新L0总地图——C2级联:K1变化需通知依赖该知识的所有对象
激活后立即执行
Step 1 读取待整合碎片列表
Read: cognitive/L2_fragments/fragment_index.md
→ 筛选出所有「🔲 待整合」和「⚠️ 部分整合」的条目
→ 如果没有待整合碎片:「当前没有待整合的碎片。知识体系已是最新状态。」→ 结束
→ 展示待整合列表给用户,询问:「是处理全部(N个),还是指定某个?」
Step 0.5 判断是否启用独立整合器(CS-012 修复,在 Step 1 之后执行)
统计本次用户选中的待整合碎片数量 N:
IF N ≥ 5:
调用 cognitive-fragment-integrator 子智能体:
输入:{
target_l1_doc_path: [目标 L1 文档路径],
fragment_ids: [选中的碎片 ID 列表],
fragment_paths: [对应的碎片文件路径列表]
}
等待 integration_plan 输出
向用户展示整合方案:
「━━ 批量整合方案(共 N 条碎片)━━
[每条碎片:fragment_id / section_name / position_type / confidence]
⚠️ confidence="低" 的碎片已标注,建议优先审查
─────────────────────
[逐条确认(推荐)] [全部确认] [逐条审查]」
用户确认后,按 integration_plan 执行写入:
a. 按方案执行 StrReplace(section_name + anchor_keyword 定位)
b. 追加变更记录(每个文档一条)
c. 更新碎片整合索引(所有已整合碎片状态改为 ✅)
d. 更新 L0 大脑总地图
e. 追加 L3 系统日志
对 unresolved_fragments(有矛盾,无法自动处理):
告知用户:「以下碎片存在矛盾,需手动决策:[碎片ID + 矛盾描述]」
→ 写入完成后跳到 Step 7(调用 cognitive-verifier)
IF N < 5:
继续原有流程(Step 2 → Step 7)
Step 2 对每个待整合碎片,用内置 explore 子智能体并行读取:
- 碎片完整内容
- 碎片关联的L1文档相关章节(只读相关章节,不是整篇)
【为什么用 explore 子智能体】
碎片内容 + L1 文档相关章节可并行读取,比串行快;
且碎片内容留在子智能体 context 中,不膨胀主对话。
主 Agent 只接收「关键摘要 + 关联段落」用于分析。
[L1.5 模式匹配]
→ 碎片是否是已有L1.5原则(P1/P2)的一个新实例?
→ 是 → 记录「印证了P?」,不额外整合进L1(已有原则的例证不需要单独进L1)
→ 否 → 继续判断
[同化/顺应判断]
→ 比较碎片与L1文档相关段落:
→ 同化(supplement):碎片是对现有内容的补充/延伸,无冲突
→ 生成补充建议:在[文档]的[章节]追加「...」
→ 顺应(revise):碎片与现有内容有张力或矛盾
→ 先运行矛盾检测(见 Step 3),再生成修订建议
→ 已覆盖(covered):现有文档已包含该观点
→ 标记碎片为「已覆盖」,告知用户,无需整合
Step 3 [矛盾检测](仅当发现顺应情况时执行)
→ 描述矛盾点:「碎片说X,但文档[Y]第Z章说了W,两者有[直接冲突/隐式张力]」
→ 基于L1.5原则推导消解方案:
选项A:修改文档[Y]的[位置]
选项B:两者都对,在文档中明确区分适用场景(添加前提条件)
选项C:碎片观点需要修正(告知用户)
→ 向用户展示,请用户决策
Step 4 生成整合建议,逐个向用户展示确认(diff 格式)
「━━ 碎片 F-XXX 整合建议 ━━
目标文档:[文档名] > [章节]
操作类型:[追加 | 修订 | 已覆盖]
┌ 建议内容 ┐
[具体的新增/修改文字]
└──────────┘
归因:🟢 AI整理(基于用户原始碎片)
风险:[🟢 低 | 🟡 中 | 🔴 高]
─────────────────────────
[✅ 确认整合] [✏️ 修改后整合] [❌ 跳过] [⏸️ 延迟]」
Step 5 执行用户确认的整合
对每个「确认」或「修改后确认」的建议:
a. Write:精确修改L1文档对应位置(追加/替换/插入)
b. 追加 [文档名]_变更记录.md(格式见下方)
c. 更新碎片整合索引.md(✅已整合 + 整合时间)
d. 更新 L0_大脑总地图.md(该文档最后更新时间)
e. 追加 cognitive/L3_logs/system_log.md
对「跳过」的:更新索引为「⚠️ 部分整合(用户跳过)」
对「延迟」的:保持 🔲 待整合 状态
Step 6 收尾汇总 + 矛盾检测(E1A 修复:顺应型整合强制触发)
「━━ 整合完成 ━━
✅ 已整合:N 个碎片
⚠️ 跳过:M 个
🔲 仍待处理:K 个
涉及文档:[文档名1]、[文档名2]」
[强制条件] 本次整合中是否包含任何「顺应(revise)」类型的操作?
→ 是(本次有顺应操作):
【强制执行】自动触发 cognitive-detect-contradiction
→ 检查范围:本次 revise 操作涉及的 L1 文档(限定范围,不做全库扫描)
→ 矛盾检测结果汇总到本步骤的输出中,用户无需再次手动确认触发
→ 否(本次全为同化/已覆盖):
告知:「本次整合均为追加操作,无矛盾风险,可跳过矛盾检测。」
(可选:「[仍然检查] [完成]」)
Step 7 调用 cognitive-verifier 子智能体(CS-010 修复)
⚠️ B4 任务日志写入在本步骤之后执行。
输入:{
target_doc_path: [本次整合涉及的主要 L1 文档路径],
update_summary: "碎片整合:共整合 N 个碎片,涉及 [章节] 的 [操作类型]",
related_docs: [本次整合中修改过的其他 L1 文档路径列表],
call_context: "integration"
}
处理验证报告(同 cognitive-update-knowledge Step 8 的处理逻辑):
→ verdict = "通过":告知用户验证通过,B4 写入(状态:完成)
→ verdict = "警告":展示警告,询问用户是否接受,接受则 B4 写入(完成)
→ verdict = "不通过":展示问题 + 建议,B4 写入(状态:挂起),任务挂起
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.
- 11d ago First seen · 260 lines · 73 tokens per session scan A 41ea4e16ee95
cognitive-integrate-fragments is a skill published in the GitHub repository TashanGKD/cognitive-os (9 stars, last pushed 5mo ago), licensed MIT. It adds 73 tokens to every session and 3,716 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. It is 88% identical to cognitive-integrate-fragments, differing in 10 lines, and is treated as a copy.
Other skills, from other repositories
cognitive-ask
A question-answering workflow that answers from a user's own knowledge documents, with sources, confidence levels, contradictions, and gaps made clear.
cognitive-reorganize
A workflow for reorganising scattered documents into a complete personal knowledge structure.
cognitive-extract-principle
A workflow for finding shared patterns across several stored thought fragments and turning them into candidate principles for review.
cognitive-integrate-fragments
A workflow for moving pending thought fragments into broader knowledge documents while preserving the existing structure.
cognitive-self-reflect
A guided self-reflection process that turns vague observations about your habits or feelings into structured records and compares them with earlier entries.
cognitive-capture-fragment
A workflow for capturing brief ideas and storing them as structured entries in a personal knowledge system.