Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/swt-labs/vibe-better-with-claude-code-vbw/llms.txt

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

All configuration lives in .vbw-planning/config.json. Settings are created during /vbw:init and backfilled automatically when new defaults are added in plugin updates.

Changing settings

Use /vbw:config to read or update any setting:
/vbw:config                          # interactive settings menu
/vbw:config effort thorough          # set a single key
/vbw:config autonomy pure-vibe
/vbw:config worktree_isolation on
/vbw:config debug_logging true
Run /vbw:config with no arguments to see all current values, the active model profile, and skill-hook mappings in a single view.

Settings reference

Every key in config/defaults.json, in order.
KeyDefaultDescription
effort"balanced"Planning and execution depth
autonomy"standard"How often agents stop to ask
auto_committrueAuto-commit after each task
planning_tracking"manual"How planning artifacts are tracked
auto_push"never"Auto-push after commits (never / after_phase / always)
verification_tier"standard"QA depth override
skill_suggestionstrueShow skill suggestions on init
auto_install_skillsfalseAuto-install detected skills
discovery_questionstrueAsk discovery questions during scope
context_compilertrueCompile role-specific context blocks for agents
visual_format"unicode"Statusline rendering (unicode / ascii)
max_tasks_per_plan5Max tasks per plan
prefer_teams"auto"When to use Agent Teams (auto / always / never)
branch_per_milestonefalseCreate a git branch per milestone
plain_summarytrueAppend plain-language phase summaries after QA
active_profile"default"Active work profile (default / prototype / production / yolo / custom)
custom_profiles{}User-defined named profiles (created via /vbw:profile save)
model_profile"quality"Model routing profile (quality / balanced / budget)
model_overrides{}Per-agent model overrides (e.g. {"dev": "opus", "qa": "haiku"})
worktree_isolation"off"Git worktree per Dev agent (off / on)
lease_lockstrueFile-level locking for concurrent agents
token_budgetstrueEnforce per-agent token budgets
two_phase_completiontrueTwo-phase task completion verification
metricstrueTrack and display session metrics
smart_routingtrueSkip unnecessary agents based on effort level
validation_gatestrueEnforce dynamic validation checkpoints
snapshot_resumetrueSave snapshots for crash recovery
event_recoverytrueRecover from interrupted events via event log
monorepo_routingtrueRoute commands in monorepo subdirectories
rolling_summaryfalseMaintain a rolling cross-phase context digest
require_phase_discussionfalseRequire discussion before planning begins
auto_uatfalseAutomatically run UAT after QA
statusline_hide_limitsfalseSuppress token-limit line in statusline unconditionally
statusline_hide_limits_for_api_keyfalseSuppress limit line only when using API key auth
statusline_hide_agent_in_tmuxfalseSuppress agent progress line in tmux sessions
statusline_collapse_agent_in_tmuxfalseCollapse full statusline to a single line in tmux worktree panes
debug_loggingfalseEnable verbose debug output (also: VBW_DEBUG=1 env var)
bash_guardtrueBlock destructive bash commands
bash_guard is not stored in defaults.json — it is read directly from project config with a default of true when absent.

Effort & Autonomy

Control how deeply agents plan and how often they stop to ask.

Model Routing

Choose which Claude model each agent uses and set turn limits.

Concurrency Controls

Manage parallel Dev agents with worktrees, lease locks, and depends_on.

Commands: Supporting

Full reference for /vbw:config and /vbw:profile.

Build docs developers (and LLMs) love