Skill v1.0.0

version: "1.0.0" name: config description: Show or change VBW settings - the model profile (quality, balanced, budget), per-role models, how much VBW does on its own, and the status line. argument-hint: "[what to change]" allowed-tools: Bash("${CLAUDE_PLUGIN_ROOT}/bin/vbw" ) Bash(vbw )

"${CLAUDE_PLUGIN_ROOT}/bin/vbw" config 2>&1 || true
"${CLAUDE_PLUGIN_ROOT}/bin/vbw" statusline status 2>&1 || true
"${CLAUDE_PLUGIN_ROOT}/bin/vbw" interview 2>&1 || true

Write to the user's level, explanation depth and involvement (above; "-": plain words).

The user said: $ARGUMENTS

Output above says not a VBW project: set up first (vbw init and vbw statusline on, one line saying so), then run vbw config again.

Change asked: make it with vbw config set KEY VALUE, show the result. Else show the settings above in plain words. Keys:

• profile: models VBW's agents run on. quality (Opus for Architect, Lead, Dev and Debugger; Sonnet for QA, Scout and Docs), balanced (default: Opus for Architect and Lead, Sonnet for the rest), budget (Haiku for Scout, Sonnet for the rest, QA included).
• model.<role> (architect, lead, dev, qa, scout, debugger, docs): override one agent (opus, sonnet, haiku, a model id; default removes it). QA never runs on Haiku, only on Sonnet or stronger (it judges whether work meets the spec): model.qa accepts sonnet, opus, default and non-Haiku model ids, and refuses any Haiku id. A stored Haiku value is raised to Sonnet.
• autonomy: how much /vbw:vibe does alone: guided (explains each step, waits), balanced (default: stops for decisions, approval, checking, shipping), hands-off (takes its own recommendations, lists them).
• autonomy_cap: steps one autonomous run takes before stopping.
• rigor: how thorough each phase's checking is. auto (default) picks a tier per phase from its risk; express, standard or deep forces it. Set with vbw config rigor MODE; phases not yet started are re-tiered.
• motion: panel animation, full, calm or off; default: by interview level.
• check_jobs: how many checks run at the same time, shared by every VBW command in this project, 1 to 64; default: three quarters of the processor cores (at least 1). Kept in this clone, not in the record. The computer-wide limit (three quarters of the cores, shared by every VBW project, not a setting; docs/proof.md) caps it.
• command_timeout: seconds vbw prove lets one project command (the test command above all) run before stopping it, 1 to 86400; default: 3600 (60 minutes). A check keeps its own timeout.
• check_budget: seconds one unit of a proof (a check, project command or test file) may take before the summaries name it over budget, 1 to 86400; default: 120. Reported only, never a limit.
• Status line: vbw statusline on or off.

Your own session keeps the model you chose with /model; it must be Sonnet, Opus or Fable for auto mode.