popup-studio-ai/bkit-claude-code

Audit shows audit logs, decision traces, and session history for AI workflows. Use it to inspect recent actions, trace a feature’s decisions, summarize activity, or search logs by action, feature, or date.

BKIT provides Claude Code commands for PDCA planning, development phases, quality review, deployment, and project setup. Use it to list available functions or drive a full feature workflow.

BKit Evals runs a skill's eval suite through evals/runner.js, validates the skill name, captures output, and saves structured results. Use it to list eval-enabled skills or check a skill's evaluation output and pass/fail counts.

BKit Explore lists installed bkit skills, agents, and evals from the local filesystem. Use it to see what is available, filter by category or level, and check eval coverage.

Bkend Auth covers bk end.ai signup, login, token refresh, RBAC, and session management. Use it when building or integrating authentication flows and access control with the bkend API.

Bkend Cookbook provides project tutorials for bkend.ai apps, from todo and blog builds to SaaS starters. Use it for implementation patterns and common error troubleshooting.

Bkend Data manages bkend.ai tables, fields, indexes, and record CRUD with filtering, sorting, relations, and pagination. Use it when working with database schemas or querying and editing table records.

Bkend Storage explains how to upload and download files with presigned URLs, CDN links, visibility levels, and multipart uploads. Use it when integrating bkend.ai file storage into code or agent workflows.

Bkit Rules defines core PDCA workflow, project level detection, agent auto-triggering, and code quality checks for Claude Code. Use it to apply bkit’s rules automatically during feature work, fixes, and reviews.

Bkit Templates provides PDCA and sprint document templates for feature planning, design, analysis, and reports. Use it when generating standardized project docs and sprint-level templates.

By-The-Way captures improvement ideas and suggestions during work without breaking task flow. Use it to record, list, analyze, promote, and track ideas for later skill creation.

CC Version Analysis examines Claude Code CLI version changes and how they affect bkit. Use it when planning an upgrade to research changes, map impacts, and produce a Korean report.

Claude Code Learning helps you configure Claude Code settings, commands, hooks, and workflows. Use it to learn features by level, generate project settings, or review newer capabilities.

Code Review analyzes files, directories, or PRs for bugs, security issues, performance problems, and code quality concerns. Use it when you need actionable review feedback on changes or staged code.

Control shows and changes bkit automation levels, trust score, and guardrails. Use it to pause, resume, or adjust how much autonomy the system has during PDCA and destructive operations.

Deploy generates CI/CD files and guides feature promotion to dev, staging, or prod with environment gates. It also supports status checks and rollback for failed releases.

Desktop App Development guides building cross-platform desktop apps with web technologies using Electron or Tauri. Use it when creating Windows, macOS, or Linux apps from a single codebase.

Development Pipeline guides a 9-phase project flow from schema and conventions through mockups, APIs, UI, review, SEO, and deployment. Use it when starting a new project or following /pipeline-* commands.

Dynamic helps set up and build full-stack web apps with bkend.ai BaaS, including authentication, database access, and API integration. Use it for Next.js projects that need login, data, and deployment patterns.

Enterprise initializes monorepo-based microservices projects with Kubernetes, Terraform, and AI-native development patterns. Use it for large-scale systems that need infrastructure templates, architecture guidance, and release planning.

Mobile App Development provides guidance for building iOS and Android apps with cross-platform frameworks like React Native, Expo, and Flutter. Use it when planning app structure, navigation, styling, API integration, or authentication.

PDCA Batch manages multiple PDCA features as independent parallel cycles with status, planning, and an interactive dashboard. Use it when features do not share scope, budget, or timeline; use /sprint for cohesive initiatives.

PDCA Fast Track auto-approves Checkpoints 1–8 when fastTrack is enabled, trust is at least 80, and a design doc exists. It falls back to L2 and manual review when any prerequisite fails.

Pdca Watch renders a read-only live PDCA dashboard that refreshes every 30 seconds from pdca-status.json and token-ledger.ndjson. Use it to monitor feature phase, match rate, iteration count, and token usage.

bkend Quick Start covers MCP setup for Claude Code, the bkend resource hierarchy, and the tenant/user model. Use it when onboarding, connecting to bkend, or creating a first project.