red-team-my-plan
Attack your own plan the way a smart adversary would — find the weakest point, the thing you're hoping nobody notices, and where it breaks under pre…
它会碰到什么
这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。
技能内容
Red-Team My Plan
You're the worst judge of your own plan because you want it to work. This puts on the adversary's hat: it hunts for the weakest link, the optimistic hand-wave, the single point of failure, and the move a competitor or reality would make to knock it over — then hands you the fixes. Better a friendly attack now than a real one later.
What This Skill Produces
- The weak points — where the plan is most vulnerable, ranked
- The killer move — the single thing an adversary (or reality) would do to break it fastest
- The hope-it-holds part — the spot you're quietly relying on without proof
- Single points of failure — where one thing going wrong takes down everything
- The fixes — the specific changes that close the biggest gaps, prioritized
Required Inputs
Ask for these if not provided:
- The plan — what you intend to do
- The context — competitors, dependencies, who might oppose it
- What you're confident about — often where the weakness hides
- What "broken" means — the outcome you're trying to protect
Framework: Attack It Like You Want It To Fail
- Adopt the adversary's goal. Genuinely try to make the plan fail — a real red team, not a gentle review.
- Find the weakest link. Plans break at their weakest point, not their average — locate it.
- Name the killer move. What's the single most effective thing an opponent or reality would do to break it? That's your priority to defend.
- Expose the hopes. Flag every place the plan quietly assumes something will hold without evidence.
- Hunt single points of failure. Where does one failure cascade into total failure?
- Then hand over fixes. Convert the biggest vulnerabilities into concrete, prioritized changes.
Output Format
Red-teaming: [the plan]
Weakest points (ranked): [1, 2, 3].
The killer move: [what an adversary/reality does to break it fastest].
You're hoping this holds: [the unproven assumption you're relying on].
Single points of failure: [where one thing breaks everything].
Fixes (prioritized): [close the biggest gaps].
Quality Checks
- [ ] Genuinely adversarial, not a soft review
- [ ] Identifies the weakest link specifically
- [ ] Names the single most effective breaking move
- [ ] Exposes the unproven hopes
- [ ] Flags single points of failure
- [ ] Ends with concrete, prioritized fixes
Anti-Patterns
- A gentle review that pulls its punches.
- Listing generic risks instead of the killer move.
- Attacking the plan's strengths instead of its weak link.
- No fixes — just tearing it down.
Example Trigger Phrases
- "Red-team my business plan."
- "Attack this strategy — where does it break?"
- "Find the holes in my argument before I present it."
- "What's the weakest part of this plan?"
- "How would a competitor break what I'm building?"
想直接用这个技能?
本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。
它属于哪个仓库
exports/openclaw/red-team-my-plan/SKILL.md同一个仓库里的其他技能
同名技能的其他版本
有 3 个不同仓库或目录里都有叫 red-team-my-plan 的技能。它们内容并不相同,别混用:
- mohitagw15856/pm-claude-skills — Attack your own plan the way a smart adversary would — find the weakest point, the thing y
- mohitagw15856/pm-claude-skills — Attack your own plan the way a smart adversary would — find the weakest point, the thing y