cognitive-calibrate

cognitive-calibrate is a skill for Cursor from TashanGKD/cognitive-os. It costs 107 tokens per session (2,082 once invoked), scanned A, original, MIT.

A confidence-review process for checking whether AI-generated or inferred knowledge was later confirmed by real evidence.

In plain words
What is it for?
Use it to find unverified content in knowledge documents and fragments, ask whether evidence now exists, and record follow-up items.
Why use it?
It helps prevent guesses from gradually being treated as facts without verification.

Skill for Cursor

Written for Cursor: installed under .cursor/.

Good fit Use it to find unverified content in knowledge documents and fragments, ask whether evidence now exists, and record follow-up items.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/tashangkd/cognitive-os/cognitive-calibrate
Install

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.

Any agent
npx skills add TashanGKD/cognitive-os --skill cognitive-calibrate
Clone the repo
git clone --depth 1 https://github.com/TashanGKD/cognitive-os

Made for: Cursor.

Wrote 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.

agentmods badge for cognitive-calibrate

README.md
[![agentmods](https://agentmods.dev/badge/skills/tashangkd/cognitive-os/cognitive-calibrate/github.svg)](https://agentmods.dev/skills/tashangkd/cognitive-os/cognitive-calibrate)
Your own site
<a href="https://agentmods.dev/skills/tashangkd/cognitive-os/cognitive-calibrate"><img src="https://agentmods.dev/badge/skills/tashangkd/cognitive-os/cognitive-calibrate/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.

agentmods 80×15 button for cognitive-calibrate

Your own site · 80×15
<a href="https://agentmods.dev/skills/tashangkd/cognitive-os/cognitive-calibrate"><img src="https://agentmods.dev/badge/skills/tashangkd/cognitive-os/cognitive-calibrate.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 107 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,082 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00107 $0.02082
Opus 5 $0.00053 $0.01041
Sonnet 5 $0.00021 $0.00416
Haiku 4.5 $0.00011 $0.00208

Measured 10d ago against content hash 46ce270e384a, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

cognitive-calibrate 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 10d 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.

.cursor/skills/cognitive-calibrate/SKILL.md · 155 lines

How it starts

The opening of the file, as written. The whole thing — 155 lines — stays where its author put it; the contents beside it link to each section on GitHub.

cognitive-calibrate(置信度校准 / 元认知置信度监控)

对应认知活动:元认知的置信度校准——追踪🟡AI生成内容是否事后得到验证,防止「自我确认偏误」导致未验证的推断长期被当作事实

认知五维坐标(COG-TAX)

  • 意识程度:显性(需要用户参与确认)
  • 脑网络:CEN(目标导向的检索与判断)+ SN(不确定性检测)
  • 记忆系统:语义记忆的准确性评估
  • 执行功能:监控(Monitoring),是元认知的核心成分
  • 双系统:系统2(需要主动评估,不能自动化)

理论依据:Flavell (1979) 元认知监控;Dunning & Kruger (1999) 元认知准确性;Nelson & Narens (1990) 置信度监控框架

调用时机

  1. 被 cognitive-consistency-check 内部调用(月度维护,Step 8)
  2. 用户主动触发(「验证历史内容」「校准知识置信度」)

Headless调用说明(被 cognitive-consistency-check Step 8 调用时):

  • 调用方传入默认范围参数:scope="ALL_90_DAYS"(跳过Step 0的交互选择)
  • 跳过 Step 0:不询问用户,直接使用 scope="ALL_90_DAYS" 执行 Step 1
  • 若筛选结果 > 10条:Headless模式下只处理前10条(避免过长执行)
  • 所有 Step 4 的结果写入 todo.md,等待下次 daily-briefing 展示给用户
  • 不向对话输出 Step 4 的摘要(Headless模式静默执行)

知识导航表

层级 文档 用途
D0 碎片索引 cognitive/L2_fragments/fragment_index.md 找到所有「归因=🟡AI生成/推断,验证状态=未验证」的条目
D0 L1 文档 cognitive/L1_knowledge/[各维度文档](含🟡标注的段落) 找到需要验证的具体内容

激活后立即执行

Step 0  确认检查范围
        Ask: 本次校准的范围?
        选项 A:90天以上未验证的🟡内容(全量,可能较多)
        选项 B:指定某个L1文档(精准,快速)
        选项 C:从上次校准以来的所有🟡内容

Step 1  读取待验证条目
        Read: cognitive/L2_fragments/fragment_index.md
        → 筛选:归因类型=🟡AI生成/推断 AND 验证状态=🔲未验证 AND capture_time < 今天-[范围天数]
        → 提取:碎片ID、标题、capture_time

        Read: cognitive/L1_knowledge/[目标文档]
        → 扫描🟡归因标注的段落
        → 提取:段落位置、内容摘要(前150字)、标注时间

        → 若筛选结果为0条 → 告知用户「选定范围内无需校准的内容(🟡内容已全部处理,或该范围内无🟡归因内容)」
           然后退出,不继续执行 Step 2 以后的步骤

Step 2  逐条向用户展示并询问

        对每个待验证条目,展示:
        ---
        📋 [内容编号/N]
        **来源**:[L1文档名 §章节] / [L2碎片ID「标题」]
        **当时的归因**:🟡 AI生成/推断([capture_date],距今[N]天)
        **内容摘要**:[前150字]

        **问题**:这段内容当时是AI的推断。现在:
        A. ✅ 已有实际证据支持(标记为「已验证」)
        B. ❌ 事后发现是错的(标记为「已否定」,需要修订L1/L2)
        C. ⏸️ 还不确定,继续观察
        D. ⏭️ 跳过这条(稍后处理)
        ---

Step 3  根据用户回答更新验证状态

        A(已验证):
        → 更新 fragment_index.md:验证状态=✅已验证,verified_date=今天
        → 更新 L1 文档:将🟡标注改为✅(StrReplace)

        B(已否定):
        → 更新 fragment_index.md:验证状态=❌已否定,verified_date=今天
        → 在 cognitive/L3_logs/todo.md 追加:
          □ [校准-YYYYMMDD] 碎片[ID]/[L1段落]已被否定,需要修订对应内容
          建议:运行 cognitive-update-knowledge 或 cognitive-detect-contradiction
        → 不自动修改 L1(修订需要用户确认)

        C(继续观察):
        → 更新 fragment_index.md:验证状态=⏸️观察中,last_check=今天
        → 不做其他操作

        D(跳过):
        → 不更新验证状态
        → 记录「已跳过」供下次校准时再展示

Step 4  生成校准摘要

        输出:
        ---
        ## 置信度校准摘要([今日日期])
        处理条目:N条
        - ✅ 已验证:X条
        - ❌ 已否定(需修订):Y条 → 已加入待完成清单
        - ⏸️ 继续观察:Z条
        - ⏭️ 跳过:W条(下次校准时再显示)
        ---

Step 5  追加系统日志
        Write: cognitive/L3_logs/system_log.md(追加)
        格式:[LOG-YYYYMMDD-NN] cognitive-calibrate | 校准完成:验证X条,否定Y条,观察Z条 | fragment_index.md

Read the full file on GitHub · 155 lines

Changes

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.

  1. 10d ago First seen · 155 lines · 107 tokens per session scan A 46ce270e384a

Subscribe to this mod's changes

cognitive-calibrate is a skill published in the GitHub repository TashanGKD/cognitive-os (9 stars, last pushed 5mo ago), licensed MIT. It adds 107 tokens to every session and 2,082 once invoked, about $0.0005 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.