pe-menezes/quick

Description and examples

What it does: For small, well-defined tasks (e.g. bug fix, tiny feature), skips PRD and full spec: generates a minimal spec in memory and a prompt pack you can hand to the coding agent. Task should fit in ≤4 files.

Examples:

• /vibeflow:quick corrigir formatação de data no dashboard — One command; you get a prompt pack and can paste it to the agent.
• /vibeflow:quick adicionar botão de exportar CSV na tela de relatórios — Same; use when scope is clear and small.

---

Language

Detect the language of the user's input ($ARGUMENTS or conversation). Write ALL output in that same language. Technical terms in English are acceptable regardless of the detected language.

Web Search Policy

Use WebSearch and WebFetch only when local context (.vibeflow/, codebase files, git history) is insufficient. Prefer local knowledge first: patterns, conventions, and existing code. Typical valid uses: researching unknown frameworks/libraries found in the codebase, checking official docs for unfamiliar APIs.

Fast-track prompt pack for: $ARGUMENTS

When to use

• Quick fixes or small features with clear requirements.
• You want a prompt pack NOW, not a paper trail.
• The task fits in ≤4 files.

When NOT to use

• The idea is vague → use /vibeflow:discover first.
• You need full documentation for the team → use the full pipeline.
• The task is large or architecturally significant → use /vibeflow:gen-spec.

Phase 0: Check context

1. Does .vibeflow/ exist?
   • YES → skip to Phase 2 (use existing knowledge).
   • NO → go to Phase 1 (lightweight scan).

Phase 1: Lightweight scan (only if no .vibeflow/)

This is NOT a full analyze. Do NOT generate .vibeflow/. Do just enough to understand the project:

1. Read project config files: package.json, pyproject.toml, Cargo.toml, go.mod, build.gradle, pom.xml, or equivalent. Determine the stack.
2. Read top-level directory structure (2 levels deep). Identify structural units.
3. Read 3-4 key files: the main entry point, one route/handler, one model/type definition, and one test file (if present).
4. If .cursorrules, CLAUDE.md, or .cursor/rules/ exist, read them for coding conventions.

Keep all findings in memory — do NOT write files.

At the end, suggest: "For deeper analysis, run /vibeflow:analyze."

Phase 2: Generate ephemeral spec

Using .vibeflow/ (if available) or Phase 1 context, generate a spec in memory only (do NOT save to file). The spec must contain:

• Objective — 1 sentence. What changes for the user.
• Definition of Done — 3-5 binary checks (fewer than standard specs).
• Scope — What's in. Keep it tight.
• Anti-scope — What's explicitly OUT. Be aggressive.
• Budget — ≤ 4 files (tighter than standard ≤6). If the task clearly needs more than 4, warn (in the user's detected language): "This task may be too large for quick. Consider using /vibeflow:gen-spec."
• Applicable Patterns — Which patterns from .vibeflow/patterns/ apply (if .vibeflow/ exists).

Do NOT include Technical Decisions or Risks sections (this is fast-track).

Phase 3: Generate prompt pack

Using the ephemeral spec and .vibeflow/ knowledge (if available), generate the prompt pack. Follow the same structure as /vibeflow:prompt-pack:

The prompt pack MUST start with:

You are only seeing this prompt; there is no context outside it.

(Write this opening line in the user's detected language.)

Then include, in this order:

1. Objective and Definition of Done

From the ephemeral spec.

2. Anti-scope

What NOT to do.

3. Budget

≤ 4 files (default for quick).

4. Patterns to Follow

If .vibeflow/ exists: embed real code examples from pattern docs and conventions.md, just like /vibeflow:prompt-pack does. If .vibeflow/ does NOT exist: include the conventions and patterns you observed during the Phase 1 lightweight scan.

5. Where to Work

Real file paths. Verify they exist. Include relevant code snippets.

6. Directional Guidance

Architectural direction. NOT step-by-step.

7. How to Run/Test (MANDATORY)

Detect test runner from stack. Always include test commands. If no test runner detected: "No test runner detected. Add manual tests to validate."

Save the prompt pack to: .vibeflow/prompt-packs/<feature-slug>.md Create .vibeflow/prompt-packs/ if it doesn't exist.

After saving, report to the user:

• Path of the generated prompt pack
• Summary of objective and DoD (2-3 lines)
• Budget used (≤ 4 files)
• If .vibeflow/ didn't exist: remind to run /vibeflow:analyze for richer results next time
• Suggest: "After implementing, run /vibeflow:audit to verify."

---

Maintenance

If this command is modified, update MANUAL.md to reflect the changes.