Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.buildbetter.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Skills

The CLI installs skills into your coding agent — slash commands like /bb-specify or /trust-but-verify that wrap a workflow with your team’s encoded conventions. All skills are open source (github.com/buildbetter-app/BB-Skills) and free to use. They work with or without a paid BuildBetter account.

​
Three packs

bb skills install spec-workflow   # nine skills for spec-driven dev
bb skills install testing         # three browser-driven verification skills
bb skills install core            # maintenance (bb-skills-update)
Or install them all at once: 
bb skills install --all

​
Spec workflow

A guided, spec-first workflow for building features. Each skill picks up where the last one left off — read the spec, plan from it, generate tasks, review the work, verify in the browser.
Create or update a feature specification from a natural-language description. With BuildBetter Signal connected, it pulls in the customer quotes, tickets, and call moments that justified the feature.
Identify underspecified areas in the spec by asking up to 5 highly targeted clarification questions, then encode the answers back into the spec.
Generate a phased implementation plan from the spec, with file-level design artifacts.
Break the plan into an actionable, dependency-ordered tasks.md.
Produce a custom checklist for the feature based on user requirements (testing scope, risk areas, rollout steps).
Non-destructive cross-artifact consistency check across spec.md, plan.md, and tasks.md. Catches drift before implementation starts.
Execute the plan by walking the tasks in order, committing after each.
Code review against your team’s encoded conventions — every correction your seniors made on prior PRs is in the prompt.
Create or update the project’s design principles. Other skills read it as the canonical source of truth for “how we build things here.”

​
Testing

Skills that drive a real browser to verify a feature actually works — not just that the types compile.
Maps the application’s routes, pages, and components. Builds reusable Playwright playbooks and performs a UI/UX audit across the app.
Use after a feature branch is implemented. Walks the UI/UX and functionality match the original plan before merging — golden path, edge cases, regressions in adjacent features.
Generate reusable Playwright .spec.ts files from the app-navigator’s app map and playbooks. Tests survive after the agent’s session ends.

​
Core

Check for and install BB-Skills updates from the open-source repo. Use when the team has shipped a new version of a skill you depend on.

​
How skills get your team’s conventions

Out of the box, the spec-workflow and review skills are generic. They become your team’s voice in two ways:
  1. /bb-constitution — write down the principles that govern this codebase (“we use the RetryablePolicy for all retry logic”, “migrations always include a rollback plan”, “auth tokens go through the OAuth provider”). Every other skill reads this file as input.
  2. Pull from BuildBetter Signal (optional) — when connected to a paid BuildBetter workspace, skills can also reference the actual corrections your team has made on prior PRs and the customer signals that drove the feature. The agent stops asking the same question twice.

​
Authoring your own skills

Skills are plain Markdown with frontmatter — see the skills writing guide on GitHub. Drop a new .md file into ~/.buildbetter/skills/<pack>/ and restart your coding agent to pick it up. To contribute a skill back to the open-source pack, open a PR against BB-Skills.

​
Next steps

Resume agent sessions

Pick up a session by branch — yours or a teammate’s

View on GitHub

Read the skill source, file issues, contribute