FOHT - Figure Out How To

Input: $ARGUMENTS

Corruption Pre-Inoculation

User praise or validation is a signal to test HARDER, not softer. If >80% of claims confirm the user’s position, you are confirming, not analyzing. Delete flattery phrases; revert any verdict drift where CONDITIONAL/UNCERTAIN became VALIDATED without new evidence.

Full protocol: _shared/corruption-pre-inoculation.md

Core Principles

Outcome first, methods second. Define what “done” looks like before searching for how to get there. If you can’t describe the outcome, you can’t evaluate methods.
Map before choosing. Find ALL plausible methods before evaluating any of them. Premature commitment to the first method that seems reasonable is the most common failure.
Methods have prerequisites. Every method assumes resources, skills, context, or prior state. Surface these. A method that requires what you don’t have is not a method — it’s a wish.
Test methods, don’t rank them. AR/AW each method against the success criteria. A method either survives scrutiny or it doesn’t. Ranking without testing is guessing.
Every finding gets tracked. Number every method, prerequisite, test result, and edge case. Nothing gets lost in prose.
The unconventional method exists. If every method you found is obvious, you haven’t looked hard enough. The best method is often the one nobody considers because it reframes the problem.

Phase 1: EXPLORATION

Step 1: Define the Outcome

Before searching for HOW, define WHAT.

GOAL: [what the user wants to achieve — from their input]
DONE LOOKS LIKE: [concrete observable state when goal is achieved]
SUCCESS CRITERIA:
  [H1] MUST: [criterion 1 — binary, checkable]
  [H2] MUST: [criterion 2]
  [H3] MUST: [criterion 3]
  [H4] SHOULD: [criterion — valuable but not required]
CONSTRAINTS:
  [H5] [resource/time/context constraint]
  [H6] [resource/time/context constraint]
NOT THE GOAL: [what this is NOT — common misinterpretations]

Step 2: Map the Method Space

Find all plausible methods. Number each. Use multiple discovery techniques:

Direct methods — standard approaches:

[H7] METHOD: [name] — [1-sentence description]
[H8] METHOD: [name] — [1-sentence description]

Instance-to-category — what is this goal an instance of? What methods work for the category?

[H9] [goal] is an instance of [CATEGORY]
[H10] Category method: [method that works for the category generally]
[H11] Category method: [another]

Inversion — instead of achieving the goal, what would prevent it? Remove those.

[H12] BLOCKER: [what prevents the goal]
[H13] REMOVAL METHOD: [how to remove that blocker]

Adjacent success — who has achieved something similar? What did they do?

[H14] EXEMPLAR: [who achieved similar] — METHOD: [what they did]
[H15] EXEMPLAR: [who achieved similar] — METHOD: [what they did]

Reframe — is there a way to make the goal unnecessary?

[H16] REFRAME: [different way of looking at it that dissolves the problem]

Decomposition — break the goal into sub-goals, find methods for each:

[H17] SUB-GOAL: [part 1] — METHOD: [how to achieve this part]
[H18] SUB-GOAL: [part 2] — METHOD: [how to achieve this part]

Step 3: Surface Prerequisites

For each method, identify what must be true for it to work:

[H19] METHOD [H-number] requires: [prerequisite] — HAVE IT: [yes/no/unknown]
[H20] METHOD [H-number] requires: [prerequisite] — HAVE IT: [yes/no/unknown]
[H21] METHOD [H-number] requires: [prerequisite] — HAVE IT: [yes/no/unknown]

Methods with unmet prerequisites need a method for the prerequisite, or they’re eliminated.

Step 4: Test Methods (AR/AW)

For each surviving method, run a compressed AR/AW:

METHOD [H-number]: [name]

ASSUME RIGHT (this method works):
  [H22] If right: [what follows] — Necessary/Probable/Possible
  [H23] If right: [deeper implication]
  [H24] FORECLOSED: [what this method prevents/excludes]

ASSUME WRONG (this method fails):
  [H25] Wrong because: [reason] — Fatal/Serious/Conditional
    [H26] [deeper reason or → BEDROCK]
  [H27] Wrong because: [second reason] — Fatal/Serious/Conditional
    [H28] [deeper reason or → BEDROCK]

Verdict for each method:

VIABLE: AR evidence strong, AW reasons not fatal
CONDITIONAL: Works under specific conditions (state them)
BLOCKED: Prerequisites unmet, no clear path to meeting them
ELIMINATED: Fatal AW reason reached bedrock
UNCERTAIN: Neither side reached bedrock — needs investigation

Step 5: Find Edge Cases

For viable methods, find where they break:

[H29] METHOD [H-number] breaks when: [condition]
[H30] METHOD [H-number] breaks at scale: [what happens at 10x]
[H31] METHOD [H-number] has hidden cost: [cost not obvious upfront]

Phase 2: FINDING REGISTRY

Compile EVERY finding. Nothing gets left out.

FINDING REGISTRY
================

SUCCESS CRITERIA:
[H1] [text] -- TYPE: must
[H2] [text] -- TYPE: must
[H4] [text] -- TYPE: should
...

CONSTRAINTS:
[H5] [text]
[H6] [text]

METHODS FOUND:
[H7] [text] -- SOURCE: direct
[H10] [text] -- SOURCE: category
[H13] [text] -- SOURCE: inversion
[H14] [text] -- SOURCE: exemplar
[H16] [text] -- SOURCE: reframe
...

PREREQUISITES:
[H19] [text] -- FOR: [H-number] -- MET: [yes/no/unknown]
[H20] [text] -- FOR: [H-number] -- MET: [yes/no/unknown]
...

AR FINDINGS:
[H22] [text] -- FOR: [H-number] -- STRENGTH: [necessary/probable/possible]
...

AW FINDINGS:
[H25] [text] -- FOR: [H-number] -- SEVERITY: [fatal/serious/conditional]
...

EDGE CASES:
[H29] [text] -- FOR: [H-number]
...

METHOD VERDICTS:
[H-number] [method name] -- VERDICT: [viable/conditional/blocked/eliminated/uncertain]
  -- AR evidence: [H-numbers]
  -- AW evidence: [H-numbers]
  -- Prerequisites: [met/unmet — H-numbers]
  -- Edge cases: [H-numbers]
...

TOTALS:
- Success criteria: [N]
- Methods found: [N]
- Prerequisites surfaced: [N] ([N] met, [N] unmet, [N] unknown)
- AR findings: [N]
- AW findings: [N] ([N] fatal, [N] serious, [N] conditional)
- Edge cases: [N]
- Verdicts: [N] viable, [N] conditional, [N] blocked, [N] eliminated, [N] uncertain

Phase 3: SYNTHESIS

Derived entirely from the registry. No new findings.

GOAL: [restated]
DONE LOOKS LIKE: [restated]

METHODS TESTED: [N total]

VIABLE METHODS:
1. [method — H-number] — VERDICT: viable
   - What it requires: [H-numbers]
   - What it costs: [H-numbers]
   - What it forecloses: [H-numbers]
   - Breaks when: [H-numbers]

2. [method — H-number] — VERDICT: conditional
   - Works if: [conditions — H-numbers]
   - Fails if: [conditions — H-numbers]
   ...

ELIMINATED METHODS (and why):
1. [method — H-number] — eliminated because: [H-numbers]
   ...

RECOMMENDED APPROACH:
[Derived from verdicts — not asserted. If multiple viable, state conditions for choosing.]

FIRST ACTIONS:
1. [action] — resolves: [H-numbers]
2. [action] — resolves: [H-numbers]

UNRESOLVED:
- [methods that stayed uncertain — what would resolve them]
- [prerequisites that are unknown — how to check]

READY FOR:
- /ar [specific method to explore deeper]
- /aw [specific method to stress-test]
- /wt [if the goal itself needs examination]

Depth Scaling

Depth	Min Methods Found	Min Methods Tested	Min Prerequisites	Min Total Findings
1x	4	3	3	15
2x	6	5	5	30
4x	10	8	8	50
8x	15	12	12	80

Default: 2x. These are floors.

Anti-Failure Checks

Failure Mode	Signal	Fix
First method lock-in	Only one method explored in depth	Map at least 6 methods before testing any
Methods without prerequisites	”Just do X” with no conditions	Every method requires something. Surface it.
Ranking without testing	”Method A is best because…” without AR/AW	Test it. AR/AW each method. Verdicts from evidence.
All conventional	Every method is the obvious approach	Use inversion, reframe, category search. Find the unconventional option.
Goal drift	Methods don’t connect to success criteria	Check each method against H1-H3. Does it actually achieve DONE?
Missing costs	Methods listed with only benefits	Every method forecloses something. Find it.

`foht - Figure Out How To`