Skip to content

foxrun-ai/foxrun / skillwright/skills/decide

decide

Verified

Decide 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.

Decide

One router, five resolvers, six frameworks. The door check sets the speed. One resolver makes the call. Every run ends in the output contract below. To record a ruling the person already made, use work.py decision-add, not this skill.

Steps

  1. State the decision in one line. Name the options. Split facts from assumptions. Done when: the line, the options, and the two lists are written.

  2. Run the door check in frameworks/two-way-door.md. A two-way door: decide now with about 70% of the facts. A one-way door: slow down, cap the downside, and ask the person when a gate in ~/.foxrun/config/gates.md applies. Done when: the door and the speed are named.

  3. Pick one resolver by the kind of decision:

    The decision is...ResolverFile
    a problem someone broughtthe 4-stepframeworks/four-step.md
    build, cut, or automate a process, tool, or systemthe algorithmframeworks/algorithm.md
    a plan about to be committed, or a stuck choiceinversion and pre-mortemframeworks/inversion.md
    an opportunity, invite, deal, or commitmenthell yes or noframeworks/hell-yes.md
    something that runs: start it, keep it, or stop itkill criteriaframeworks/kill-criteria.md

    Done when: one resolver file is open.

  4. Answer the resolver's questions in order, one line each. Done when: each question has its line.

  5. Report with the output contract. Done when: every field is filled, and the next step is one action with an owner.

Output contract

  • Decision:
  • Door: two-way / one-way
  • Resolver:
  • Facts / assumptions:
  • Risk accepted (what breaks):
  • Next step + owner:
  • Kill criteria (only if something now runs):

Keep it to one screen, decision first, in plain words.

Chaining

  • door check -> any resolver
  • 4-step -> the algorithm, when the answer is "build or cut something"
  • inversion -> hell yes or no, when the choice is still open after the failure scan
  • any resolver -> kill criteria, when the decision starts something that runs

Gotchas

  • Chain a second resolver only when the first one names it. One resolver per decision is the default.
  • This skill never widens a gate. Money, sends, deletes, and production changes still go to the person.
  • Each framework file names its source. Cite it when the person asks why.
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: 468kUpdated Claude CodeCodexVerifiedInstall
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: 321kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 321kUpdated Claude CodeCodexVerifiedInstall
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 availableStars: 321kUpdated Claude CodeCodexVerifiedInstall