foxrun-ai/foxrun / skillwright/skills/decide
decide
VerifiedDecide one choice through a door check and one resolver: two-way door, the 4-step, the algorithm, inversion, hell yes or no, and kill criteria. Use when an agent or the person must decide, size the rigor, stop something, or say no.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 2.6 KB |
| frameworks/algorithm.md | markdown | 1.1 KB |
| frameworks/four-step.md | markdown | 1.1 KB |
| frameworks/hell-yes.md | markdown | 766 B |
| frameworks/inversion.md | markdown | 1.1 KB |
| frameworks/kill-criteria.md | markdown | 1.1 KB |
| frameworks/two-way-door.md | markdown | 1.3 KB |
| replays/read-door-first.json | config | 255 B |
| replays/read-output-contract.json | config | 201 B |
frameworks/two-way-door.md
View on GitHub# Two-way door
**Source:** Jeff Bezos. Amazon shareholder letters: 2015 (Type 1 and Type 2 decisions) and 2016 (decide with about 70% of the information you wish you had; disagree and commit). Downside capping from Richard Branson, "Losing My Virginity" (1998): he got Boeing to take the Virgin Atlantic 747 back if the airline failed in year one. Protect the downside, then chase the upside.
**Use when:** every decision. This runs first and sets the speed.
## The questions, in order
1. Can we walk back through this door cheaply? Count the reversal cost in time, money, and trust.
2. **Two-way door.** Decide now with about 70% of the info. Act. Fix it later if wrong. Do not escalate.
3. **One-way door.** Slow down. What is the worst case? Can we survive it? Can we cap it with a small test, an exit clause, or a return option? Then get more input, and escalate if an `AGENTS.md` gate applies.
4. Waiting past 70% is its own cost. Name what the delay costs.
## Output
- Door: two-way / one-way
- Reversal cost:
- Worst case + cap (one-way only):
- Speed: decide now / slow down / escalate
## Traps
- Treating every decision as one-way. Most are two-way.
- Treating a one-way door as two-way because it feels small. Money, sends, deletes, and production changes are one-way by default.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |