Stop babysitting your AI agent.
The Damper CLI gives AI agents everything they need — task specs, project architecture, code conventions, and quality gates — so they ship features that actually fit your project.
npx @damper/cli setup npx @damper/cli # connect specs, repo context, and # Damper task flow from your terminal
AI agents are powerful, but without structure they produce inconsistent, misaligned code.
Copy-pasting context into every session
You manually paste architecture docs, coding conventions, and task requirements into each AI chat. Miss something and the agent builds the wrong thing. Every session starts from zero.
Agents skip steps and ignore rules
Forgot the migration? Skipped the tests? Used the wrong naming convention? Without enforcement, AI agents cut corners you'd never accept in a code review.
Features that don't match your project
The agent writes code that works in isolation but doesn't follow your patterns, uses different error handling, or introduces dependencies you don't want. You end up rewriting half of it.
Zero copy-paste, full context
The CLI auto-injects the task spec, project architecture, code conventions, and templates into the agent's context. Every session starts fully informed — nothing to forget.
Rules that can't be skipped
Critical rules are shown when the agent starts. Completion checklists require proof — test output, build logs — before a task can be marked done. No rubber stamps.
Code that fits your project
Project context teaches agents your patterns: file structure, naming conventions, error handling, testing approach. Every feature ships consistent with the rest of your codebase.
Your code never leaves your machine
Damper is designed so that no source code or sensitive data is shared with our servers. Here's what stays private:
You define what to build. AI handles how to build it. The better your spec, the better the result.
- Add a task in Damper with a description and implementation plan. Include acceptance criteria, technical requirements, and any constraints.
- The CLI fetches your task spec, project architecture docs, code conventions, and templates — then writes it all into TASK_CONTEXT.md inside the worktree.
- Claude Code launches with full context already loaded. It reads your spec, follows your project patterns, and starts implementing.
- AI commits its work, you review. Merge when satisfied. The task is marked complete in your roadmap automatically.
Four steps from task to shipped code.
- Run setup to store your API key and configure the Damper MCP server for Claude Code. Takes 30 seconds.
- Add a task in the Damper dashboard with a description and implementation plan. The more detail you provide, the better AI performs.
- Pick a task, and the CLI creates a worktree, injects context, and launches Claude Code. AI starts coding immediately.
- Review the AI's commits, merge the worktree back to your branch, and clean up. Task is marked complete in your roadmap.
Stop babysitting your AI agent.
Run one command and let AI code your tasks. The Damper CLI creates isolated worktrees, injects full task context, and launches Claude Code automatically.