Claude Launchpad

/lp-enhance

AI-powered instruction-file improver — runs as a /lp-enhance skill inside Claude Code or Cursor Agent.

After init, the instruction file has the basics but not the details. The agent still guesses your architecture, conventions, and off-limits zones. /lp-enhance runs inside the session, reads the codebase, and fills in what matters — CLAUDE.md on Claude Code, AGENTS.md on Cursor Agent.

/lp-enhance

Installation

init offers installation only when /lp-enhance is not already installed (project/global/legacy). In interactive mode you can choose project scope, global scope, or skip. With --yes, it defaults to project scope.

  • Claude project scope (.claude/skills/lp-enhance/SKILL.md) — this project only
  • Claude global scope (~/.claude/skills/lp-enhance/SKILL.md) — all Claude Code projects
  • Cursor project scope (.cursor/skills/lp-enhance/SKILL.md) — written by init --harness cursor

If missing on Claude, doctor --fix creates the project-scope skill. Cursor --fix repairs the Cursor skill when Launchpad manages it.

What it does

  • Fills in Architecture, Conventions, Off-Limits, Memory & Learnings, Key Decisions (skips Backlog, which you maintain manually)
  • Stays under the 200-instruction budget — overflows detailed content to .claude/rules/ or .cursor/rules/*.mdc
  • Suggests project-specific hooks (Claude: SessionStart widened for compact/clear; Cursor: .cursor/hooks.json snippets)
  • Suggests path-scoped rules and monorepo excludes
  • Suggests MCP servers based on detected external services
  • Runs inside your session — the agent already has your codebase context

Instruction budget

CLAUDE.md and AGENTS.md have a soft limit of ~200 actionable lines. Past that, the agent starts ignoring rules at the bottom. Doctor warns at 200. The skill targets 200, and moves detailed content to harness rule files when overflow occurs.

When to re-run

The instruction file gets stale as your codebase changes. Re-run /lp-enhance when:

  • You've done a major refactor or added new modules
  • Your architecture has shifted (new services, changed patterns)
  • The agent keeps making wrong assumptions about your codebase
  • Doctor shows instruction sections are missing or incomplete
  • It's been a few sprints and the instructions no longer match reality

Next

On this page