VividKit Crystal clear AI coding
VI
Choose Your Setup Get Started with AgentKit Continue with ClaudeKit Migrate from ClaudeKit to AgentKit Clean cutover from ClaudeKit What Is AgentKit? How AgentKit Works AgentKit Kits, Skills, Agents & Hooks AgentKit Runtime Support AgentKit Desktop App AgentKit Configuration AgentKit CLI Quick Reference AgentKit Skills Cheatsheet AgentKit Workflows AgentKit Statusline Update AgentKit CLI & Kits AgentKit Helper Clean reinstall of AgentKit AgentKit Troubleshooting What is ClaudeKit? ClaudeKit CLI Guide Command Finder Commands Cheat Sheet Workflow Recipes CLI Quick Reference Claude Mechanics Permission Modes ClaudeKit Hooks Coexistence ClaudeKit x Codex CK → Other coding tools How ClaudeKit Works Inside ClaudeKit Finding Your Unknowns Session Recovery Fix from Logs UI/UX Guide Accounts Switcher IDE Config Remote Control Codex App System Diagramming UI Review Gate Promotions Support VividKit

Quick Ref / Command

Engineer Kit

/ak:sequential-thinking

Reflective step-by-step analysis

Use reflective step-by-step analysis with dynamic expansion, contraction, revision, branching, hypothesis verification, and explicit final convergence.

01

Estimate loosely

02

Structure thoughts

03

Adjust dynamically

04

Mark revisions

Rule 01

The total number of thoughts is adjustable, not a promise

Rule 02

Revision is a feature, not a failure

Rule 03

Branches must converge with explicit rationale

Rule 04

Use visible markers only when useful; otherwise apply the method internally

Execution Map

Core lanes from input to output, aligned with the command's real execution path.

01

Start

Input, scope, route

  1. 1 Estimate loosely Start with an initial thought count such as Thought 1/5, knowing the total can change.
  2. 2 Structure thoughts Each thought builds on prior context, handles one aspect, states assumptions or uncertainty, and points to the next step.
02

Work

Agent / skill execution

  1. 3 Adjust dynamically Expand when complexity appears, contract when simpler, revise when insight invalidates earlier work, and branch for alternatives.
  2. 4 Mark revisions When correcting a previous thought, name the original, why it changed, and what downstream impact follows.
03

Verify

Gate, review, validation

  1. 5 Explore branches For competing approaches, label branches from the parent thought, compare them explicitly, then converge with rationale.
  2. 6 Verify hypotheses Generate hypotheses, run or reason through verification evidence, and iterate until the candidate solution is supported.
04

Close

Report, handoff, artifact

  1. 7 Finalize only ready Mark final only when the solution is verified, critical aspects are covered, confidence is adequate, and no key uncertainty remains.

Syntax, arguments, subcommands

How to invoke the skill: syntax, positional arguments, shared options, then each subcommand with its own syntax and outcome.

Syntax

/ak:sequential-thinking [problem to analyze step-by-step]

Arguments

[problem to analyze step-by-step] Required

Problem to analyze

States the concrete decision, diagnosis, constraints, available evidence, rejected alternatives, authority boundary, and useful response bound for revisable step-by-step reasoning.

/ak:sequential-thinking "Choose a migration strategy for splitting the audit log from the primary database. Use at most 8 concise checkpoints, compare dual-write and change-data-capture branches, revise assumptions when evidence conflicts, and finish with a decision, verification plan, and unresolved questions."

Modes / Routes

Explicit

Visible thought markers

Research:
Use for requested breakdowns
Review:
None
Validate:
Final marker after verification
Implicit

Internal reasoning aid

Research:
Use for routine problem solving
Review:
None
Validate:
Cleaner user-facing answer

Sample Prompt

Concrete invocations for each flag, subcommand, mode, or route available in the live workflow.

Architecture tradeoff Recommended 
/ak:sequential-thinking compare event sourcing vs audit-log tables for this service
Use when:
Use when a complex decision has dependent tradeoffs and may need branching or revision.
Expected:
Starts with a loose thought estimate, structures assumptions and alternatives, branches the competing approaches, verifies the strongest option, and finishes with a final decision rationale.
Debug hypothesis 
/ak:sequential-thinking analyze why uploads intermittently fail after deploy
Use when:
Use for unclear failures where hypotheses need verification and course correction.
Expected:
Builds a thought sequence around likely causes, marks revised assumptions when evidence changes the model, records verification steps, and stops only when the supported cause is clear.
Adaptive implementation plan 
/ak:sequential-thinking plan a safe migration from REST polling to webhook delivery
Use when:
Use when scope may expand or contract as risks, dependencies, and rollout constraints become clearer.
Expected:
Produces a step-by-step plan that can adjust its thought count, names uncertainties and next areas to inspect, revises invalidated steps, and ends with verified readiness criteria.
Branch alternatives 
/ak:sequential-thinking choose between queue-backed retries, cron repair, and synchronous retry for payments
Use when:
Use when several approaches deserve separate exploration before convergence.
Expected:
Labels branches from a parent thought, compares the alternatives explicitly, captures any revisions from new insights, verifies the preferred path, and marks a final thought only when uncertainties are resolved.

Handled Scope

  • Complex decomposition
  • Adaptive planning
  • Revision cascades
  • Alternative comparison
  • Hypothesis-driven debugging
  • Uncertainty management

Next