FAQ

Does BotMinter only work with Claude Code?

Claude Code is the recommended backend and all shipped profiles are built for it today. Under the hood, BotMinter uses Ralph orchestrator, which also supports Gemini CLI, Codex, Kiro, Amp, Copilot CLI, and OpenCode. Profiles for these backends are planned.

Do I need multiple agents running in parallel?

Not at all. It depends on the profile you choose. The agentic-sdlc-minimal profile uses three roles (engineer, chief-of-staff, sentinel) with the engineer wearing multiple hats — planner, implementer, reviewer. Other profiles like scrum distribute those hats across more specialized agents. You pick the formation that fits your workflow.

Is this only for Scrum teams?

No. BotMinter focuses on conventions, not a specific methodology. It ships with opinionated defaults (like the Scrum profile), but profiles are fully customizable. Think of it like Rails for web or Spring for enterprise — baked-in conventions that you can tweak to fit your process. You can fork an existing profile or create one from scratch.

What's the difference between BotMinter and just using Claude Code?

Claude Code is a powerful single-agent tool. BotMinter adds structured conventions on top: layered knowledge scoping (team → project → member), process enforcement via status gates, cumulative learning through knowledge files, and a coordination fabric via GitHub issues. The value isn't in the runtime — it's in the conventions that make your investment in one agent carry forward to every other agent you run.

What is a "profile"?

A profile is a reusable methodology template — it defines roles, pipeline stages, quality gates, and knowledge structure. bm init extracts a profile into a new team repo. You can use the built-in profiles as-is, fork them, or create your own.

What is Minty?

Minty is BotMinter's interactive assistant — a lightweight persona shell that you launch with bm minty. Unlike team members, Minty is not orchestrated by Ralph and does not run in a loop. It is simply a coding agent session primed with BotMinter-specific knowledge and skills, designed for ad-hoc operator tasks like browsing profiles, checking team status, or getting guidance on hiring decisions.

Minty works without any teams configured — if ~/.botminter/ doesn't exist, it runs in "profiles-only" mode and can still browse profiles and answer general questions.

How does Minty differ from team members?

Team members (launched with bm start or interacted with via bm chat) are full Ralph Orchestrator instances — they have hats, a persistent loop, memories, and coordinate through GitHub issues. Minty has none of that. It is a one-shot coding agent session with a persona prompt and a set of composable skills. Think of it as the difference between a team player running a structured workflow and a knowledgeable helper you can ask quick questions.

What can Minty do?

Minty ships with four composable skills:

• team-overview — reads ~/.botminter/config.yml and reports on teams, members, running state, and workspace status
• profile-browser — explores profiles at ~/.config/botminter/profiles/, showing roles, statuses, coding agents, and conventions
• hire-guide — interactive guide for bm hire decisions: which role to hire, naming suggestions, and implications
• workspace-doctor — diagnoses workspace health: checks for missing files, stale submodules, broken symlinks, and Ralph lock state

These skills are markdown files under ~/.config/botminter/minty/skills/ and are injected into the coding agent session via Minty's persona prompt.

Why do I need to prefix my comments with @bot?

Each agent has its own GitHub App identity and posts as a bot user (e.g., team-engineer[bot]). The @bot prefix helps the agent reliably identify human input when parsing issue comments. While agent and human comments are now distinguishable by author, the prefix remains a useful convention for clarity.