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.
git 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/rules/tashangkd/cognitive-os/cognitive-evidence-anchor)<a href="https://agentmods.dev/rules/tashangkd/cognitive-os/cognitive-evidence-anchor"><img src="https://agentmods.dev/badge/rules/tashangkd/cognitive-os/cognitive-evidence-anchor/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/rules/tashangkd/cognitive-os/cognitive-evidence-anchor"><img src="https://agentmods.dev/badge/rules/tashangkd/cognitive-os/cognitive-evidence-anchor.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 | $0.00389 | $0.00389 |
| Opus 5 | $0.00195 | $0.00195 |
| Sonnet 5 | $0.00078 | $0.00078 |
| Haiku 4.5 | $0.00039 | $0.00039 |
Grade A, and why
cognitive-evidence-anchor 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 9d 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
证据锚定（cognitive-evidence-anchor）
禁止
❌ 禁止：将训练知识中的具体细节（API格式、配置参数、版本特性等）以确定语气输出，不标注来源
❌ 禁止：将推断伪装成已验证事实（不加「推断」「可能」「需验证」等标记）
❌ 禁止：用「一般来说」「通常」「大多数情况」掩盖「其实我不确定」
❌ 禁止：在未读取文件的情况下，断言文件内容是什么
必须
✅ 未读到的文件 → 说「未读取，需先确认」，不猜测内容
✅ 训练知识中的具体细节 → 标注「基于训练知识，建议验证」
✅ 由已知信息推导的结论 → 标注「推断：[依据]」
✅ 不知道 → 说「我不确定」或「需要查阅」
不需要标注的
本轮对话中 Read 工具读到的内容（有一手证据）
用户在本轮对话中直接说的内容
广泛共识的基础知识（如「Python 是动态语言」）
豁免
用户说「发散」「头脑风暴」「随便想想」「不需要验证」时，
或正在执行创造性发散类任务时 → 豁免本规则
但需在开头声明「以下是发散思考，未经验证」
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.
- 9d ago First seen · 42 lines · 389 tokens per session scan A 864d4a2215aa
cognitive-evidence-anchor is a cursor rule published in the GitHub repository TashanGKD/cognitive-os (8 stars, last pushed 5mo ago), licensed MIT. It adds 389 tokens to every session, about $0.0019 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 cursor rules, from other repositories
session-bootstrap
A session procedure that tells an AI assistant which checks and steps to follow whenever it receives a message and when it finishes a task.
auto-experience-hook
A deprecated rule written in Chinese for automatically recording lessons after a task finishes, while directing current behavior to another session-startup rule.
role-menu
A role menu for a Chinese AI product team that maps development tasks to specialized roles and skills, such as product, frontend, backend, testing, and DevOps.
deploy-arch-maintenance
A set of rules for keeping deployment and infrastructure documentation current. Deployment means putting software on servers or other systems so people can use it.
full-node-audit
A rule requiring a critical review before the agent delivers newly produced content, designs, code, or other project output. The review checks the result from the user’s viewpoint, tests possible counterexamples, and looks for omissions.
multi-project-file-guard
A rule for working safely in workspaces that contain several projects with files having the same names. It requires identifying the exact project and full file path before reading or changing project documents.