Take ideas to implementation with a planning-command pipeline
A solo SaaS founder uses a custom five-command agent flow (intent, impact analysis, planning, execution, review) with lint guardrails to ship features in a production product.
Done withClaude Code
- What they did
- The founder built five chained commands (intent, impact analysis, iterative planning, execution orchestration, review/demo) that all update one plan file in git, which he reviews at each stage. A shared CONTRIBUTING.md guides agents, ESLint rules run as post-edit hooks, and a Bash-approval script screens commands. When the agent misbehaves, it writes a new rule with tests.
- How it went
- An orchestrator script turning plans into workflows cut token spend to about 20% and runtime from 5-6 hours to 20-30 minutes. He still reviews most code via git diff, and calls it mostly clean.
- Worth knowing
- The guardrails grew incrementally: each unwanted agent action or approval prompt became a new lint rule or approval rule, so expect to keep investing in them.
Try it yourself with Claude Code
Set up a planning pipeline for my project in [repo or product name] as custom slash commands: one to capture intent, one for impact analysis, one for planning, one for execution, and one for review. Make the execution step run [my lint and test commands] and stop if they fail. Test it on this small feature: [feature idea]. Stop when the feature has gone through all five steps and I have a reviewed diff, and do not commit or push without asking me.
Source: HN · Sep 6, 2026
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Turn tickets into merged pull requests
Pulls tasks from GitHub Issues or Linear, then runs Claude Code or Codex in isolated Kubernetes pods to carry each one through to a PR.
Run content research and sponsor outreach with Claude computer use
The author tested Claude computer use in Cowork and Claude Code over two days on six workflows including research, sponsor outreach and product listings.
Build and balance a live Telegram RPG
A non-coding PM had Claude Code write a Telegram RPG bot, deploy it over SSH, and run tens of thousands of test fights to rebalance classes.
- Claude
Build a C compiler with a team of parallel agents
Anthropic's engineering write-up on tasking Opus 4.6 agent teams with building a C compiler.
Build and test a simple endpoint project with Devin
A reviewer had Devin set up a project, write code, install dependencies, test endpoints and add error handling in about 4 minutes on a simple task.
Build a booking and payment app for tattoo artists
A former programmer used Replit Agent over a weekend to build a booking app with calendar, Google OAuth, Stripe deposits and notifications, though time zones and UI needed back-and-forth.