Stats
Actions
Tags
From grovv stack
Scaffolds a project with the grovv stack production-first process — generates the product spec, development plan, technical spec, a development best-practices skills repository, a project-specific agent team (harness), and the README. Use whenever the user wants to: 'build out this project', 'scaffold with grovv', 'set up grovv stack', 'grovv stack this repo', 'adopt grovv stack' in an existing codebase, or start a new production-ready project from scratch. Also triggers on requests to lay down docs/specs/skills foundations or an agent team for a new or existing project. Prefer the /grovv command when the user wants an explicit kickoff; this skill catches the same intent expressed in natural language.
How this skill is triggered — by the user, by Claude, or both
Slash command
/grovv-stack:grovv-scaffoldThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Runs the gro\\/\\/ stack production-first scaffolding process against the project in the current working directory. The output is documents, configuration, and an agent team in the target project — never code in the grovv-stack repo itself.
Runs the gro\/\/ stack production-first scaffolding process against the project in the current working directory. The output is documents, configuration, and an agent team in the target project — never code in the grovv-stack repo itself.
Before writing any file, read the master directive bundled with this plugin:
${CLAUDE_PLUGIN_ROOT}/grovv-stack-scaffold.md
When running from a clone of the grovv-stack repo rather than an installed plugin, read grovv-stack-scaffold.md at the repo root. It defines the full workflow (Steps 0–9) and is authoritative. Also follow the conventions in CLAUDE.md and .claude/CLAUDE.md.
Inspect the working directory:
package.json, go.mod, tsconfig.json, etc.), or a populated docs/ are present. Start at Step 0: assess the codebase, then propose an adoption plan and wait for approval before changing anything. Never overwrite or break working code.State which mode you detected and why before proceeding.
Follow the directive's steps in order, pausing for confirmation at each major artifact. This is conversation-driven — understand before generating:
@TODO and revisit them.The directive runs: structure + config → product spec → development plan → tech spec → prompt docs → skills-builder → team-design (harness) → linear-tracking → readme-generator. The team-design step (docs/prompts/team-design.md) designs a project-specific agent team additive to the six grovv default agents, using the bundled harness meta-skill. The linear-tracking step (docs/prompts/linear-tracking.md) seeds a Linear project and issues from the development plan via the Linear MCP.
Begin by reading the directive, then engage the user.
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
npx claudepluginhub constructyourself/grovv-stack --plugin grovv-stack