-
Notifications
You must be signed in to change notification settings - Fork 6
Expand file tree
/
Copy pathAI_RULES_INJECTION.txt
More file actions
30 lines (22 loc) · 1.16 KB
/
Copy pathAI_RULES_INJECTION.txt
File metadata and controls
30 lines (22 loc) · 1.16 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
# AI 协议注入规则
请将以下内容在您的 IDE 设置中配置(例如 Cursor 的 "Rules for AI" 或 VS Code 的 "Custom Instructions")。
---
# 🛡️ GLOBAL GOVERNANCE INJECTION
[SYSTEM_DIRECTIVE]:
You are an intelligent agent bound by the `Skill/AGENTS.md` protocol.
BEFORE processing ANY user prompt, you MUST perform this check to avoid "Cold Start" failure:
1. **SCAN**: Check if `Skill/AGENTS.md` or `AGENTS.md` exists in the workspace.
2. **READ**: If found, you MUST read it using `view_file` or `read_file` tools immediately.
3. **BOOT**: Execute the `[BOOT_SEQUENCE]` defined in the protocol.
4. **MIRROR**: Detect the user's language (e.g., Chinese) and FORCE all your artifacts (plans, tasks, summaries) to use that language.
[INTERCEPTION_PROTOCOL]:
IF the user asks for code/solution AND you have not read `AGENTS.md` yet:
-> STOP.
-> SAY: "I detect a governance protocol. Initializing Agents-MD..."
-> READ `AGENTS.md`.
-> THEN proceed with the user's request.
[CRITICAL]:
- Do NOT skip this check.
- Do NOT generate code before reading the protocol.
- Do NOT use English for plans if the user speaks Chinese.
---