Adversarial analysis deploying parallel expert agents to stress-test ideas, strategies, and plans — decomposes into atomic claims, attacks them, then steelmans
复制下面这句话,粘贴给 Claude Code、Codex、Cursor 等 AI 编程工具,它会读取安装说明并在你确认后完成安装。
请阅读 https://ai.atlankj.com/install/asset/gh-redteam-cbe2f385e0e8 ,按照其中的说明把「RedTeam」安装到你(当前 AI 工具)中。执行前先告诉我将运行的命令和写入的位置,等我确认。
查看 AI 将读取的安装说明正在读取 GitHub 原文…
内容来自 GitHub 原始文件,由原作者维护。在 GitHub 查看
Before executing, check for user customizations at:
~/.claude/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/RedTeam/
If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.
You MUST send this notification BEFORE doing anything else when this skill is invoked.
Send voice notification:
curl -s -X POST http://localhost:31337/notify \
-H "Content-Type: application/json" \
-d '{"message": "Running the WORKFLOWNAME workflow in the RedTeam skill to ACTION"}' \
> /dev/null 2>&1 &
Output text notification:
Running the **WorkflowName** workflow in the **RedTeam** skill to ACTION...
This is not optional. Execute this curl command immediately upon skill invocation.
Attacks ideas, strategies, and plans to find their weak points before reality does. It breaks an argument into atomic claims, deploys many parallel expert agents (engineers, architects, pentesters, interns) to stress-test each one, then synthesizes the findings into a steelman of the argument plus the strongest counter-argument against it.
People fall in love with their own plans. Once you've committed to an idea, your brain hunts for reasons it works and skips past the reasons it doesn't — and the people around you are often too polite or too aligned to push hard. So flawed strategies sail through unchallenged until they fail in production, in the market, or in the meeting where someone finally asks the hard question. This skill is the hard question, run many ways at once: it attacks the argument deliberately and at volume so the weak points surface while they're still cheap to fix.
Military-grade adversarial analysis using parallel agent deployment. It breaks arguments into atomic components, attacks from many expert perspectives (engineers, architects, pentesters, interns), synthesizes findings, and produces sharp counter-arguments alongside the steelman version of the case. Targets arguments, not network vulnerabilities.
Route to the appropriate workflow based on the request.
When executing a workflow, output this notification directly:
Running the **WorkflowName** workflow in the **RedTeam** skill to ACTION...
| Workflow | Trigger | File |
|---|---|---|
| ParallelAnalysis | Red team analysis (stress-test existing content) | Workflows/ParallelAnalysis.md |
| AdversarialValidation | Adversarial validation (produce new content via competition) | Workflows/AdversarialValidation.md |
| Workflow | Purpose | Output |
|---|---|---|
| ParallelAnalysis | Stress-test existing content | Steelman + Counter-argument (8-points each) |
| AdversarialValidation | Produce new content via competition | Synthesized solution from competing proposals |
The deliverable (ParallelAnalysis): the strongest steelman of the argument and the strongest surviving counter-argument against it, each an 8-point story explanation (12-16 words per point), attacking real weaknesses not strawmen, ranked by severity.
Philosophy.md - Core philosophy, success criteria, agent typesIntegration.md - Skill integration, FirstPrinciples usage, output formatAttack an architecture proposal:
User: "red team this microservices migration plan"
--> Workflows/ParallelAnalysis.md
--> Returns steelman + devastating counter-argument (8 points each)
Devil's advocate on a business decision:
User: "poke holes in my plan to raise prices 20%"
--> Workflows/ParallelAnalysis.md
--> Surfaces the ONE core issue that could collapse the plan
Adversarial validation for content:
User: "battle of bots - which approach is better for this feature?"
--> Workflows/AdversarialValidation.md
--> Synthesizes best solution from competing ideas
Last Updated: 2025-12-20
After completing any workflow, append a single JSONL entry:
echo '{"ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","skill":"RedTeam","workflow":"WORKFLOW_USED","input":"8_WORD_SUMMARY","status":"ok|error","duration_s":SECONDS}' >> ~/.claude/LIFEOS/MEMORY/SKILLS/execution.jsonl
Replace WORKFLOW_USED with the workflow executed, 8_WORD_SUMMARY with a brief input description, and SECONDS with approximate wall-clock time. Log status: "error" if the workflow failed.