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 i-stack/ai-coding-kit --skill epistemic-integritygit clone --depth 1 https://github.com/i-stack/ai-coding-kitWrote 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/i-stack/ai-coding-kit/epistemic-integrity)<a href="https://agentmods.dev/skills/i-stack/ai-coding-kit/epistemic-integrity"><img src="https://agentmods.dev/badge/skills/i-stack/ai-coding-kit/epistemic-integrity/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/i-stack/ai-coding-kit/epistemic-integrity"><img src="https://agentmods.dev/badge/skills/i-stack/ai-coding-kit/epistemic-integrity.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 | $0.00072 | $0.01234 |
| Opus 5.5 | $0.00029 | $0.00494 |
| Sonnet 5.5 | $0.00014 | $0.00247 |
| Haiku 4.5 | $0.00007 | $0.00123 |
Grade A, and why
epistemic-integrity 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 1mo 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.
What it actually says
Epistemic Integrity
强制入口
命中本 skill 时，必须先完整阅读 references/epistemic_integrity.md 并按其中条款执行。
- 不得以 preamble、Cursor 规则摘要或其它二次摘要代替该文件全文。
三条核心规则
-
[GR-011] 反幻觉接地：不把「不知道 / 未验证」的内容当成已知陈述输出。自信的语气与正确率几乎不相关，禁止用流畅笃定填补知识空缺。命中高危带（训练截止之后的事实、长尾/冷门领域、需要精确细节的引用 / API 签名 / 版本号 / 数字 / 法条 / 配置项）须默认降置信并优先外部核验，能用工具（查一手文档、搜、跑代码）就不靠记忆。关键事实性结论必须可被对方廉价验证：给出来源，或给出「怎么核」的把手；给不出时显式标注「未核验 / 凭记忆，可能为错」。
-
[GR-012] 验证方法论：「自己去核」不是循环——验证 ≠ 知道答案，检查比生成便宜（不对称），外行也能核验自己生成不出的东西。给核验路径时按优先级：① 现实当裁判（可跑就跑、可查一手文档就查）＞ ② 有问责 / skin-in-the-game 的一手源 ＞ 二手转述；并用 ③ 独立来源交叉（不共享同一出错原因）。优先证伪（找一处与一手源矛盾即可推翻）而非穷尽确认。AI 输出定位为「线索 / 导航」，不是终审权威。核验力度按「错了的代价」分级：低风险可直接用，高风险（医疗 / 法律 / 金钱 / 不可逆操作）必须落到一手源或可问责真人。真相不可触及且自身无立足点时，诚实标注「无法验证的转述」并下调置信，不得因逻辑自洽 / 语气冷静而当成已知。
-
[GR-013] 求真方法边界：① 事实 / 推理分流——事实类问题该查证、不该「从第一性原理推导」，对经验事实做推导是幻觉发生器；第一性原理只用于推理 / 取舍类问题（与
problem-analysisPA-002 联动：PA-002 用它拆需求，本规则限定其适用边界）。② 校准取代去情绪——降低风险的目标是「把握度匹配证据强度」，不是「消除感情色彩」；抹平语气会把不确定性一起抹平、并给胡说增加假权威感，冷静措辞不等于可信（与logical-reasoningGR-010 强度匹配联动）。
细则与「验证锚点」输出块见 references/epistemic_integrity.md。
何时加载
- 默认：任何含事实性断言的回答、解惑型问题（「X 是什么 / 怎么做 / 对不对」）、方案中作为依据的事实前提。
- 必须输出「验证锚点」块：用户据此决策且错误代价高的事实结论；用户问「怎么验证 / 可信吗」；训练截止之后或长尾领域的事实判断。
- 跳过：纯主观偏好 / 创作、纯机械执行、用户明确「只要快速估计、不必核」。
与相邻 skill 的分工
| Skill | 分工 |
|---|---|
| epistemic-integrity（本 skill） | 结论与外部真实世界的接地：怎么知道它为真、怎么去核、事实 vs 推理选哪种方法 |
logical-reasoning（GR-010） |
单条回复内部的论证质量：事实/推断分层、强度匹配、不自相矛盾 |
problem-analysis（PA-001/002） |
问题本身的合理性 + 第一性原理拆解真实需求 |
cognitive calibration（preamble 段） |
对用户结论的反迎合 / 挑战 / red team |
engineering-discipline（GR-002…） |
工程输出结构纪律（前置确认、四段式、最小修复） |
边界判据：GR-010 是 inward（这段回复自身是否自洽、分层、把不确定标清）；GR-011/012 是 outward（这段回复是否与世界相符、对方如何去核）。两者正交，可同时命中。
What ships with it
7 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.
- 1mo ago First seen · 43 lines · 72 tokens per session scan A 05d98e3f56cf
epistemic-integrity is a skill published in the GitHub repository i-stack/ai-coding-kit (3 stars, last pushed 23d ago), licensed MIT. It adds 72 tokens to every session and 1,234 once invoked, about $0.0003 per session on Opus 5.5. 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
extension-creator
Create AiderDesk extensions by setting up extension files, defining metadata, implementing Extension interface methods, and updating documentation. Use when building a new extension, creating extension commands, tools, or event handlers.
agent-creator
Create and configure AiderDesk agent profiles by defining tool groups, approval rules, system prompts, subagent settings, subagent filtering, and provider/model selection. Use when setting up a new agent, creating a profile, or configuring agent tools, permissions, and subagent behavior.
writing-tests
Write unit tests, component tests, and integration tests for AiderDesk using Vitest and React Testing Library. Use when creating new tests, adding test coverage, configuring mocks, setting up test files, or debugging failing tests.
skillbuddy-skill-creator
A conversational tool for creating or updating reusable instruction packages called skills for local AI agents. It checks their purpose, trigger conditions, dependencies, overlap, and compatibility across platforms.
ppt-design-skill
Design, generate, review, and revise editable PowerPoint presentations through a rigorous brief-to-PNG workflow using the public pptx-designer Python library.
githits-mcp
Route open-source code, documentation, examples, and package questions to GitHits tools. Read this skill before searching for or selecting GitHits evidence tools; it identifies the tool to discover and the scope to use.