skills.rest
jeremybrasher
jeremybrasherCOMMUNITY
Explore all skills by this author →
GitHub

figure-it-out

Designs an auditable multi-phase playbook for large migrations and unmatched engineering tasks.

Updated Aug 28, 2026
Search Tags:#code review#verification#migration#task decomposition#audit trail#playbook design#hypothesis loop
One-click install
npx skills add https://github.com/jeremybrasher/grokbot-skills --skill figure-it-out-jeremybrasher
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: figure-it-out
Source: https://github.com/jeremybrasher/grokbot-skills/tree/main/collections/pstack/skills/figure-it-out
Command: npx skills add https://github.com/jeremybrasher/grokbot-skills --skill figure-it-out-jeremybrasher

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? When a task fits no existing playbook—like a large migration or an ambitious multi-part change—this Skill designs a bespoke, auditable workflow instead of improvising, so a human can review the work after stepping away. ## Core Features & Use Cases - Playbook Design Before Code: Frames a falsifiable done predicate, quantifies scope, sets a rigor level, and decomposes work into atomic, independently-landable units sequenced riskiest-first. - Hypothesis Loop with Audit Trail: Runs each unit as an experiment verified against the real artifact, logging every decision in a canonical TSV trail via the show-me-your-work skill. - HOLD Discipline: Stops and returns the gap when required inputs are missing, rather than guessing or treating inconclusive results as passes. - Use Case: You face a repository-wide migration with no existing playbook. The Skill frames the done predicate, builds a verification harness from the pre-change baseline, executes units under the loop, and hands back a committed decision trail for review. ## Quick Start Ask the agent to figure out a plan for migrating the legacy billing module to the new API, with a falsifiable done predicate and a decision log you can review afterward.

Frequently Asked Questions about figure-it-out

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I plan a large code migration with an AI agent?▼

Use this Skill to frame a falsifiable done predicate, decompose the migration into atomic landable units sequenced riskiest-first, and build a verification harness before writing features. Each unit runs as a hypothesis loop verified against the real artifact.

What is a falsifiable done predicate in task planning?▼

A falsifiable done predicate is a concrete, testable condition that defines when the task is complete, such as a check comparing old value versus new value. The Skill refuses to start a long run without one.

When should I not use the figure-it-out playbook?▼

Skip it when a narrower bundled playbook already fits, when the work is a standing multi-day program rather than one bespoke run, or when you cannot state any done predicate. The Skill itself routes these cases to a HOLD.

How does the audit trail work for reviewed agent work?▼

The Skill logs one canonical TSV with a row per decision and per unit, with evidence as links, via the show-me-your-work skill. The trail is committed alongside the diff so a reviewer can trust the work after stepping away.

What happens when required inputs are missing?▼

The Skill stops and returns a HOLD describing what is missing and what it needs, rather than inventing inputs. A HOLD is a return value, not a delay before guessing.

Related Skills

Explore More Software Engineering Skills→
Sma1lboySma1lboy
136

figure-it-out

Designs an auditable multi-phase playbook for large migrations and ambitious cross-cutting changes.

Community
Advanced
mhawthornemhawthorne
11

gza-code-review-interactive

Review committed changes on a feature branch against main.

Community
Advanced
adjohnadjohn
3

figure-it-out

Designs an auditable multi-phase playbook for large migrations and ambitious changes.

Community
Advanced
DuySeuDuySeu
1

gsd-fast

Execute small, one-off tasks inline without subagents or planning.

Community
Basic