vibekit

Use after verify-gate returns `ready`, as an optional security pass over a vibe run's diff before finish-branch. Runs a tiered threat-model review — universal code-security checks on every diff, plus AI-artifact checks when the diff builds a skill, agent, prompt, or MCP server — quotes verbatim file:line evidence, scores findings, and hard-gates the handoff on CRITICAL/HIGH with a written-waiver escape. Peer to review-pack; both optional, either may run before finish-branch.

Use to diagnose vibekit installation health — skill file integrity, runtime registration parity, `.vibekit/` directory health, `docs/` subdir presence, authoring-contract alignment. Reports verdict only by default; `--fix` auto-repairs safe items. Cross-runtime portable.

Use after verify-gate returns `ready` and before finish-branch or any outward-facing action. Performs a self-review pass (Reflexion-style) — the skill critiques its own completed work against the spec and the plan, surfaces issues in order of severity, and presents the diff to the user for explicit sign-off. No outward-facing action runs until the user approves.

Use when starting any conversation — establishes vibekit's auto-trigger discipline so pipeline skills (brainstorm, brief, plan, exec, verify, review, finish, memory) actually fire at their trigger points instead of being silently skipped.

Use when the user says "vibe" or gives a short intent like "add X", "build Y", "fix Z". Runs the full disciplined pipeline end-to-end — brainstorm → plan → isolated execution → verification — with minimal user interaction outside the brainstorm phase. The user only has to answer design questions; every other step is autonomous, guardrail-protected, and token-efficient. If the user does not say "vibe" explicitly, do not assume this skill applies.

You MUST use this before any creative or implementation work — creating features, building components, adding functionality, or modifying behavior. Runs a disciplined Socratic brainstorming loop (hard gate, one-question-at-a-time, 2-3 approaches, design doc, user approval) while keeping token overhead minimal. Questions, user answers, approaches, design, and the written spec are never compressed.

Use when executing an approved implementation plan task-by-task. Dispatches one fresh subagent per task using an RTCO brief, receives the return through the report filter, runs two-stage review (self + plan-compliance), and only then marks the task complete and moves on. Prevents context pollution, drift across tasks, and silent regressions.

Use to store, recall, query, or audit durable project knowledge — atomic facts and compound documents alike — plus a working notepad that survives conversation compaction. One skill, one storage convention, file-backed under `.vibekit/`. Cross-runtime portable.

Use when a spec has been approved by the user and implementation has not started. Converts the spec into a bite-sized, TDD-structured implementation plan with exact file paths, full code snippets, and explicit commit boundaries. The plan is written as normal prose (not compressed) because downstream execution agents parse it verbatim.

Use when a verification or test fails and you need a disciplined response instead of guessing — fires as vibekit's pipeline failure branch (verify-gate not-satisfied, exec-dispatch test/build failure) and standalone when the user reports a bug, failing test, or broken build. Finds and proves a root cause with verbatim evidence, then routes the fix through exec-dispatch or plan-write. Never edits code itself.

You MUST use this before any creative or implementation work — creating features, building components, adding functionality, or modifying behavior. Runs a disciplined Socratic brainstorming loop (hard gate, one-question-at-a-time, 2-3 approaches, design doc, user approval) while keeping token overhead minimal. Questions, user answers, approaches, design, and the written spec are never compressed.

Use before dispatching any subagent or task. Compiles a verbose intent into a tight RTCO (Role, Task, Constraints, Context, Output) brief that minimizes tokens while preserving every guardrail. Mandatory before any Task/Agent dispatch.

Use when executing an approved implementation plan task-by-task. Dispatches one fresh subagent per task using an RTCO brief, receives the return through the report filter, runs two-stage review (self + plan-compliance), and only then marks the task complete and moves on. Prevents context pollution, drift across tasks, and silent regressions.

Use after review-pack has returned `yes` and the user has explicitly signed off on the diff. Presents the user with concrete integration options (merge, PR, keep branch, cleanup) and executes only the one the user picks. Never auto-merges, auto-pushes, or auto-PRs. This is the outward-facing endpoint of the vibe pipeline.

Use after a plan is approved and before any implementation dispatch. Creates an isolated workspace — a git worktree (preferred) or a fresh branch — scoped to the current vibe run. Prevents in-progress work from mixing with the user's other work, makes rollback cheap, and gives exec-dispatch a clean slate to operate in.

Use to store, recall, query, or audit durable project knowledge — atomic facts and compound documents alike — plus a working notepad that survives conversation compaction. One skill, one storage convention, file-backed under `.vibekit/`. Cross-runtime portable.

Use when a spec has been approved by the user and implementation has not started. Converts the spec into a bite-sized, TDD-structured implementation plan with exact file paths, full code snippets, and explicit commit boundaries. The plan is written as normal prose (not compressed) because downstream execution agents parse it verbatim.

Use to drive a vibe run autonomously across multiple iterations until verify-gate returns `ready` — bounded by iteration / wallclock / commit budgets, halted on thrashing or genuine ambiguity. Same verification gates as a single vibe run; never bypasses review-pack sign-off. Cross-runtime portable with degraded fallback on runtimes lacking native loops.

Use immediately after receiving output from any subagent, tool chain, or dispatched task. Parses the raw return, enforces the schema declared in the brief, strips pleasantries and restatements, and surfaces only the signal. Rejects non-conforming output.

Use after verify-gate returns `ready` and before finish-branch or any outward-facing action. Performs a self-review pass (Reflexion-style) — the skill critiques its own completed work against the spec and the plan, surfaces issues in order of severity, and presents the diff to the user for explicit sign-off. No outward-facing action runs until the user approves.

Use when starting any conversation — establishes vibekit's auto-trigger discipline so pipeline skills (brainstorm, brief, plan, exec, verify, review, finish, memory) actually fire at their trigger points instead of being silently skipped.

Use before claiming a feature is done, before merging, and before running any finishing / release skill. Verifies the implementation against the spec using evidence-based checks and three independent self-consistency passes on the question "does this satisfy the spec?". Evidence (test output, diffs, error messages) is quoted verbatim. Silent passes are not allowed — every check must produce evidence or it does not count.

Use when the user says "vibe" or gives a short intent like "add X", "build Y", "fix Z". Runs the full disciplined pipeline end-to-end — brainstorm → plan → isolated execution → verification — with minimal user interaction outside the brainstorm phase. The user only has to answer design questions; every other step is autonomous, guardrail-protected, and token-efficient. If the user does not say "vibe" explicitly, do not assume this skill applies.

Use to diagnose vibekit installation health — skill file integrity, runtime registration parity, `.vibekit/` directory health, `docs/` subdir presence, authoring-contract alignment. Reports verdict only by default; `--fix` auto-repairs safe items. Cross-runtime portable.