Skip to content
FunCoding

Search

Search docs, Skills and MCP

distill-shield

Distill Shield:为个人移交资料包生成 Canary 与可选加固策略,提高未授权蒸馏成本;仅适用于有权处置的数据。

AI 与智能体1.1kdistill-shield-skill/SKILL.md

Install

Send this to Claude Code, Codex or Cursor. The agent checks the Skill for safety first and installs it only after you confirm.

读取 https://funcoding.ai/skills/agenmod/immortal-skill/distill-shield-skill/install.md ,按里面的步骤帮我安装这个 Skill。

SKILL.md

复制给 AI

  • 本 Skill 唯一入口:当前文件 distill-shield-skill/SKILL.md。
  • 仓库根:假定 immortal-skill/(或含本目录与 kit/ 的克隆根)。
  • 从仓库根运行生成脚本(推荐,避免路径歧义):
python3 distill-shield-skill/kit/shield_gen.py --output ./handover-bundle --label "我的移交包"

Distill Shield

语言

根据用户第一条消息的语言,全程使用同一语言。

何时激活

  • 用户要「防蒸馏」「加固资料包」「移交前埋 Canary」。
  • 用户运行 kit/shield_gen.py 后需要根据输出做伦理与可读性检查。

伦理红线

  • 仅对本人或已获授权的数据使用;禁止对他人资料恶意投毒。
  • 不承诺数学上不可攻破;承诺 可检测性、审计成本、部分管线干扰。

路径约定

  • 本 Skill 根目录 {baseDir} = distill-shield-skill/。

操作顺序

Phase 0:威胁模型

  • 蒸馏场景:一次性 LLM 蒸馏 vs 批量训练 vs 仅人类阅读?
  • 接收方:家人 / 律师 / 公司?(决定策略强度)

Phase 1:材料盘点

  • 格式、是否分块检索、是否经 OCR。

Phase 2:策略选择(准则)

策略对人对自动化管线可被绕过
Canary 唯一字符串低影响(可放附录)泄露时可检索剔除后丢失
提示注入段需标注阅读区可能干扰单次 LLM 蒸馏预处理可剥离
矛盾样本可能困惑读者污染蒸馏一致性人工可识别

Phase 3:生成

在 distill-shield-skill/ 目录内时:

python3 kit/shield_gen.py --output ./handover-bundle --label "我的移交包"

在 仓库根(推荐,与「复制给 AI」一致):

python3 distill-shield-skill/kit/shield_gen.py --output ./handover-bundle --label "我的移交包"

将生成的 CANARY.txt 与说明合并进实际交付目录。

Phase 4:自检清单

  • 第三方隐私已脱敏
  • Canary 已记录在安全处(便于日后比对)
  • 人类读者能看懂「附录为技术性标记」

Phase 5:(可选)给接收方的一页说明

解释「附录含技术性标记,用于完整性校验,非正文内容」。

不做的事

  • 不提供针对具体模型的对抗样本保证。
  • 不帮助绕过合法安全控制。

Similar Skills

brand-guidelines
anthropics/skills180k

brand-guidelines

Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.

AI & agents

internal-comms
anthropics/skills180k

internal-comms

A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).

AI & agents

template-skill
anthropics/skills180k

template-skill

Replace with description of the skill and when Claude should use it.

AI & agents

mcp-builder
anthropics/skills180k

mcp-builder

Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node/TypeScript (MCP SDK).

AI & agents

algorithmic-art
anthropics/skills180k

algorithmic-art

Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.

AI & agents

academy-guide
anthropics/skills180k

academy-guide

Stop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: "how do I", "how can I", "getting started with", "what can Claude do", "teach me", "learn to use"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.

AI & agents