Quality score 69strikethroo.canpicasoft.com#coding#no-code#writing#design

Strikethroo

Strikethroo is an AI task management tool that uses Agent Skills to break down tasks into plans and execute them via two commands (st-create-plan and st-execute-blueprint), with optional automated code review.

Visit website

About this tool

Strikethroo is a structured AI task management tool that leverages Agent Skills to decompose tasks into plans and execute them. It operates through two main commands: st-create-plan to draft a work plan, and st-execute-blueprint to build the task. Each command is an Agent Skill that loads automatically when you describe what you need in plain language. Task decomposition happens during execution, and an optional automated code review runs at the end. Additionally, an optional review gate runs in between: a discovered second harness critiques the cumulative diff and records its findings for you, before you read the result. The plan gets your careful read, and the task blueprint is generated for you rather than reviewed by you. You can create a plan by asking your assistant in plain language, for example: /st-create-plan create user authentication with email/password and JWT tokens. The plan is saved at ai/strikethroo/plans/01--user-authentication/plan-01--user-authentication. Optionally, you can ask a second assistant to refine the plan using the st-refine-plan skill for a two-agent feedback loop. This skill is blueprint-only and does not revisit complexity analysis. Strikethroo is built on Cloudflare's tech stack, uses a subscription-based business model, and is open-source.

Use cases
  • Create a project plan using natural language
  • Automatically decompose tasks into executable blueprints
  • Run automated code review after task execution
  • Refine plans through a two-agent feedback loop

Features & use cases

Agent Skills-based automatic task decompositionTwo core commands: st-create-plan and st-execute-blueprintOptional automated code reviewOptional review gate with a second harness for critiqueSupports two-agent feedback loop (st-refine-plan)Plans stored as plain text files

Tech stack

Cloudflare

At a glance

Monetization
  • Subscription
Maturity signals
  • Open source
Learning curve
  • Medium

Resources

Sources

Every profile keeps its sources and latest verification date.