peel/fiddle

deliver

Use after implementation to analyze drift, update documentation, evolve evaluation, and close a completed epic.

orchestrate

Use when starting or resuming a full development lifecycle for a feature or explicit epic.

write-plan

Use when you have a spec or requirements for a multi-step task, before touching code.

using-fiddle

Use when working with the Fiddle portable skills library, choosing lifecycle skills, or adapting Fiddle instructions across Claude, Codex, and Pi harnesses.

develop-holistic

Use after all per-task evaluations complete — assesses cross-domain integration and creates remediation beans

develop-loop

Use when a single task bean needs implementation and evaluation — called by fiddle:develop, not directly.

develop

Use when implementing an epic's task beans through the evaluator loop — after plan and beans exist

brainstorm

Use before creative work or behavior changes to clarify intent, explore tradeoffs, and produce an approved design.

challenge

Use when a plan or design needs assumptions, constraints, and decision branches resolved before progressing.

define

Use after discovery to turn confirmed scope into an approved design, implementation plan, and ready-to-execute beans.

adr

Use when recording a consequential technical decision; captures context, the chosen option, and consequences in a new architecture decision record.

backlog

Use when capturing an idea, technical debt, or observation that is not ready for a planned bean.

define-beans

Use when sizing a planned task into task or feature beans based on its test-driven implementation complexity.

deliver-docs

Use after a completed epic or significant change to update technical docs and capture resulting decisions or debt.

discover-docs

Use when bootstrapping or reviewing project documentation through concise Socratic discovery and targeted updates.

discover

Use before defining a feature or epic to gather context, research options, and challenge scope assumptions.

feedback

Use when recording a user signal with its context, observation, implication, and confidence for future analysis.

panel

Use when evaluating architectural approaches or design decisions — runs structured multi-model adversarial analysis with cross-review and synthesis

debug

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes

finish-branch

Use when implementation is complete, all tests pass, and you need to decide how to integrate the work - guides completion of development work by presenting structured options for merge, PR, or cleanup

init

Use when starting fiddle on a new project — scaffolds docs, orchestrate.json, and beans from templates.

insights

Use when preparing for a planning cycle or when feedback has accumulated — synthesizes structured feedback entries into persona files and insight summaries that discover and brainstorm phases consume.

tdd

Use when implementing any feature or bugfix, before writing implementation code

worktrees

Use when starting feature work that needs isolation from current workspace or before executing implementation plans - creates isolated git worktrees with smart directory selection and safety verification

verify

Use when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always

quickfix

Use when the user wants a quick one-shot implementation of a small, unambiguous task — skips full orchestrate in favor of clarify, implement, PR

evaluate

Use when scoring an implementation against its task spec — dispatched by develop-loop, not directly

runtime-evidence

Use when an evaluator needs to interact with a running application before scoring dimensions