agentteamland

/dst-handoff <prototype-name> [--target flutter|react-admin|react-public] — bundle a prototype + its linked design system into a handoff package and brief the appropriate code agent (flutter-agent / react-agent) to integrate into the project's actual source code. One command from design to code.

/dst-questions — manage the project's open product/design questions list (.atl/wiki/open-questions.md) and keep it pinned to CLAUDE.md so questions blocking screen / prototype work are impossible to miss. Modes: init (scaffold), sync (re-pin from current file), add, resolve, list.

/dst-delete-ds <name> — remove a design system from the project's .dst/. Warns if any prototype depends on this DS (and refuses without --force). Updates .dst/state.json and re-renders .dst/index.html.

/dst-delete-prototype <name> — remove a prototype from the project's .dst/. Updates .dst/state.json and re-renders .dst/index.html.

/dst-edit-ds <name> "<change>" — apply a textual change to an existing design system. Reads current ds.json, invokes ds-architect-agent to interpret the change, updates files, re-renders detail.html and .dst/index.html. Optionally re-renders linked prototypes if structural changes affect them.

/dst-edit-prototype <name> "<change>" — apply a textual change to an existing prototype. Reads current prototype.json + linked DS, invokes prototype-agent to interpret the change, updates files, re-renders preview.html and .dst/index.html.

/dst-init — bootstrap or refresh the .dst/ directory in the current project. BOOTSTRAP: creates state.json + styles.css + tailwind.min.css + empty index.html on fresh projects. REFRESH: safely re-copies templates/CSS and re-renders every HTML page from preserved JSON state, so upgrading the team package always yields up-to-date visuals without touching user content.

/dst-new-ds <name> — create a new design system in the current project's .dst/ directory. Asks an interactive set of Q&A to gather brand intent + scope, then invokes ds-architect-agent to produce ds.json + Tailwind-rendered detail.html. Updates .dst/state.json and re-renders .dst/index.html.

Move dirty changes to an auto-named branch, run save-learnings + docs-sync drain + AI review chain (generic baseline + team specialists), open PR. Optional --auto-merge enables GitHub auto-merge with polling + auto-fix loop. End-of-work returns the user to the target branch.

Drain marker-driven docs queue, run drift-detector against the changed scope, regenerate TR mirrors via parity-checker, present accept/reject drafts. Boring sessions free (pre-flight short-circuit). Invoked from /create-pr Step 4.5 OR manually for catch-up audits. Composite design from the docs-sync-automation brainstorm (Q1+Q3+Q4+Q5+Q7).

Generate a full code-level Mermaid class diagram of the entire project. Shows all classes, interfaces, relationships, dependencies, and inheritance. Output saved to .atl/docs/code-diagram.md

Persist conversation learnings to journal + wiki + agent children + skill learnings, with auto-rebuild of agent.md / skill.md Knowledge Base sections from frontmatter. Multi-transcript input via --transcripts. Updates the learning-capture state file at completion (closes the auto-trigger loop). AskUserQuestion gates only for new structure creation: new agent, new skill, new rule, agent identity change, skill core change. Everything else is automatic.

Project knowledge base — living, cross-referenced, always current. Initialize, ingest new knowledge, query existing, lint for staleness. Powered by Karpathy's LLM Wiki pattern.

Scaffold a complete production-ready project with a single command. Asks key questions (project name, SaaS?, modules), then creates the full stack: .NET API (Vertical Slice + Clean Arch), Docker Compose, PostgreSQL, RabbitMQ, Redis, Elasticsearch, logging pipeline, email pipeline, auth, and optionally Flutter/React apps.

Orchestrate the Claude Design loop for a new screen. /design-screen start "<one-line description>" → derives slug, gathers requirements, builds a token-aware prompt, writes intent.md (status: pending-design), hands prompt to user. /design-screen done <handoff-url> → finds the matching pending design (or asks if multiple), fetches the handoff bundle, extracts it, hands a self-contained brief to flutter-agent or react-agent for integration. State persists in .claude/design/{slug}/intent.md across sessions — full Phase A context survives compactions and restarts.

End-to-end system health verification. Checks all containers, ports, endpoints, and pipelines (logging, email, auth, socket, worker). Proves the system is truly working, not just running.

Comprehensive workspace status report — every peer repo's branch + clean state, every cached install path, every open PR across the agentteamland org, with stale-merged-branch warnings and one-line cleanup commands. Use when the user asks "are all repos on main", "any open PRs", "is the working tree clean", "anything pending", "what's the state", or whenever the user wants to verify nothing has drifted. Self-contained — no follow-up prompts, just runs the checks and reports.

End a session with every repo on a clean main, no open PRs, no uncommitted work. Companion to /repo-status — where status REPORTS drift, cleanup ACTS on it. Auto-runs /save-learnings first, then drains uncommitted work via PRs, auto-merges them (skill-scoped exception to the merge-discipline rule), tags releases, updates registry pinning, and prunes stale branches across all peer repos + cached install paths. Use when the user wants the whole org buttoned up — typically end-of-day, before a break, or after a flurry of half-shipped work.

A clarification wizard that uses option-based Q&A rounds before adding a rule. Captures missing details from context, presents alternatives as options, dynamically detects multiple rules, and finally adds the finalized rule(s) via /rule. 3 scopes: project, --global, --team.

Add a new coding/architecture rule. The user describes it in natural language (any language), the skill writes it in English structured format to the correct file. 3 scopes: project (default), --global, --team.

Start and complete brainstorming sessions. start = begin a new brainstorm, done = complete the active brainstorm and propagate to the document chain. 3 scopes: project (default), --global, --team.

DEPRECATED — use the atl CLI binary instead. Run atl install <team>, atl list, atl remove <team>, or atl update. The /team skill is kept here only as a stub that points users to atl.