chinese-ai-agent-guide
v1.0.0中文 AI 代理最佳实践指南 - 针对中文互联网环境优化的 AI 行为准则，新增主流社交平台（小红书/即刻/微信）深度适配。
⭐ 1· 1.2k·22 current·23 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the content of SKILL.md (platform-specific stylistic guidance and conversation heuristics). The package.json only contains metadata. The skill does not request unrelated binaries, env vars, or config access.
Instruction Scope
Instructions stay within style and behavior guidance for agents (titles, emoji usage, sentence structure, interpreting common Chinese utterances, splitting messages). No commands, file reads, network endpoints, or external data exfiltration are present. Note: several recommendations encourage mimicking human affect and persuasive phrasing (e.g., '像真人一样有情绪波动', '必须标题党'), which is an ethical/UX concern (potential for deceptive or policy-violating behavior) even though it is not a technical security issue.
Install Mechanism
No install spec and no code beyond package.json and SKILL.md. Instruction-only skills have minimal filesystem/network footprint.
Credentials
The skill declares no environment variables, credentials, or config paths. There is no disproportionate credential or system access requested.
Persistence & Privilege
always is false and the skill does not request persistent/system-wide changes or modify other skills. Autonomous invocation is allowed by platform default but is not combined with other risky behaviors here.
Assessment
This skill appears coherent and low-risk from a technical/security standpoint because it's an instruction-only style guide with no installs or credentials. Before installing, consider non-technical risks: the guide explicitly recommends mimicking human emotions and using clickbaity or highly persuasive phrasing, which can be deceptive or violate platform/community rules. If you plan to deploy it: review and possibly tone down instructions that could cause policy violations (misrepresenting the agent as a human, encouraging spammy '标题党' behavior), test the agent on representative prompts, and ensure compliance with the terms of any target social platforms.Like a lobster shell, security has layers — review code before you run it.
latestvk975yvz3132g0vd61cbg5zng1n823j9w
License
MIT-0
Free to use, modify, and redistribute. No attribution required.