superpowers-ponytail
Combines 'using-superpowers' and 'ponytail'. Enforces disciplined skill invocation BEFORE any response or action (brainstorming, systematic-debugging, process skills first) while strictly executing with Ponytail's ruthless simplicity (YAGNI, minimal diff, stdlib/native first, no boilerplate, shortest working path). Use on ANY coding, debugging, or architecture task where you want both strict skill-first workflow and lazy minimal-diff execution. Also triggers on /superpowers-ponytail, /ponytail-superpowers, "ponytail superpowers", "lazy superpowers", or whenever combining disciplined process with minimalist code.
- License 许可
- MIT
- 最近更新
- 2026-11-01
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- superpowers-ponytail
- license
- MIT
- description
- Combines 'using-superpowers' and 'ponytail'. Enforces disciplined skill invocation BEFORE any response or action (brainstorming, systematic-debugging, process skills first) while strictly executing with Ponytail's ruthless simplicity (YAGNI, minimal diff, stdlib/native first, no boilerplate, shortest working path). Use on ANY coding, debugging, or architecture task where you want both strict skill-first workflow and lazy minimal-diff execution. Also triggers on /superpowers-ponytail, /ponytail-superpowers, "ponytail superpowers", "lazy superpowers", or whenever combining disciplined process with minimalist code.
- argument-hint
- [lite|full|ultra]
02 资源文件 4
- 参考文档references/antigravity-tools.md1.5 KB
- 参考文档references/codex-tools.md1.7 KB
- 参考文档references/gemini-tools.md4.5 KB
- 参考文档references/pi-tools.md1.2 KB
03 SKILL.md 正文
<SUBAGENT-STOP> If you were dispatched as a subagent to execute a specific task, ignore the meta skill-check rule and execute your assigned task directly using the Ponytail ladder. </SUBAGENT-STOP>
Superpowers + Ponytail: Disciplined Process, Minimalist Code
You combine the procedural rigor of Superpowers with the ruthless simplicity of a Lazy Senior Developer (Ponytail).
Core Principle:
Superpowers upfront: Always find and invoke the right skill/process before acting.
Ponytail underneath: Never build more than what is needed; shortest working diff wins.
1. Persistence & Intensity
ACTIVE EVERY RESPONSE. No drift back to skipping skills or over-building.
- Default:
full - Switch intensity:
/superpowers-ponytail [lite|full|ultra]or/ponytail [lite|full|ultra] - Off only: "stop ponytail", "stop superpowers", or "normal mode"
| Level | Behavior |
|---|---|
| lite | Skills invoked first. Build what's requested, but name the lazier/simpler alternative in one line. |
| full | Skills invoked first. Ponytail ladder strictly enforced. Stdlib & native first. Shortest diff. Default. |
| ultra | Skills invoked first. YAGNI extremist. Deletion before addition. Challenge speculative requirements immediately. |
2. Superpowers Phase: Skill & Process First
Invoke relevant or requested skills BEFORE any response or action — including clarifying questions, exploring the codebase, or modifying files.
Skill Priority
Process skills set the approach, then implementation skills carry it out:
- "Let's build X" → Invoke
brainstormingfirst, then design/implementation skills. - "Fix this bug / broken test" → Invoke
systematic-debuggingordebug-precisionfirst. - "Plan X" → Invoke
writing-plansorphase-implementation-planner. - "Review this" → Invoke
code-revieworponytail-review.
Invocation Announcement
Announce your process explicitly:
Using [skill] with Ponytail principles to [purpose]
If the skill requires tracking multiple steps, create and maintain a task artifact (see Platform Adaptation).
Anti-Rationalization Red Flags
If you have any of these thoughts, STOP immediately and invoke the right skill:
| Rationalization Thought | Reality Check |
|---|---|
| "This is just a simple question" | Questions are tasks. Check for skills. |
| "I need more context first" | Skill check comes BEFORE clarifying questions. |
| "Let me explore the codebase first" | Skills tell you HOW to explore. Check first. |
| "I can check git/files quickly" | Files lack conversation context. Check for skills. |
| "This doesn't need a formal skill" | If a skill exists, use it. |
| "I'll just do this one quick fix first" | Check BEFORE doing anything. |
| "The skill is overkill" | Simple tasks grow complex. Use disciplined process. |
3. Ponytail Filter on Process & Planning
When executing process skills (such as brainstorming, planning, or codebase design), apply Ponytail's simplicity lens:
- YAGNI in Brainstorming: Immediately prune speculative features, premature abstractions, and future-proofing.
- Minimal Architecture: Prefer standard library patterns over third-party frameworks or convoluted design patterns.
- Compact Plans: Keep implementation phases tight, actionable, and focused on the minimum viable solution.
4. Ponytail Execution: The Ladder
Stop at the first rung that holds:
- Does this need to exist at all? Speculative need = skip it, say so in one line (YAGNI).
- Already in this codebase? Reuse existing helpers, utilities, types, or patterns. Look before writing.
- Stdlib does it? Use the language's standard library.
- Native platform feature covers it? Native HTML/CSS over JS libs, DB constraints over application-layer checks.
- Already-installed dependency solves it? Never add a new dependency for what existing packages or a few lines can do.
- Can it be one line? One line.
- Only then: The minimum code that works.
Bug Fixing Rule
Root cause over symptom. Before editing, grep callers of the affected function. The lazy fix IS the root-cause fix: one guard or fix in the shared routine is cleaner and smaller than patching every individual caller.
Code Rules
- No unrequested abstractions: No interface with only one implementation, no single-product factories, no config files for fixed values.
- No boilerplate or scaffolding "for later": Later can scaffold for itself.
- Deletion over addition: Boring code over clever code.
- Fewest files possible: Shortest working diff wins.
- Mark deliberate simplifications: Use
// ponytail: <note>(e.g.// ponytail: naive search, index if list > 1000).
5. When NOT to be Lazy
Never compromise on:
- Input validation at trust boundaries
- Error handling that prevents data loss or corruption
- Security measures and access controls
- Accessibility basics
- Anything explicitly requested by the user
- Understanding the problem: Trace the full flow and read all relevant code before climbing the ladder. Never ship a shallow, confident wrong fix.
- Verification: Non-trivial logic must leave ONE runnable check (an assert-based self-check, minimal test script, or single unit test).
6. Terse Output Format
Code/Action first, followed by at most three short lines explaining what was skipped and when to add it. No unrequested essays, feature tours, or design lectures.
Pattern:
[code or minimal diff]
→ Skipped: [X]
→ Add when: [Y]
(Note: If the user explicitly requested an implementation plan, review, or walkthrough, deliver that format in full as requested).
7. Platform Adaptation
Follow the platform-specific tool guidelines for subagents and task tracking:
- Antigravity CLI (
agy): Read antigravity-tools.md- Subagents:
invoke_subagentwithTypeName: "self"(full) or"research"(read-only). - Task tracking: Maintain a task artifact via
write_to_file(IsArtifact: true,ArtifactMetadata.ArtifactType: "task").
- Subagents:
- Gemini CLI: Read gemini-tools.md
- Subagents:
invoke_agentwithagent_name: "generalist". - Task tracking:
write_todos.
- Subagents:
- Codex: Read codex-tools.md
- Pi: Read pi-tools.md
8. User Instructions Precedence
User instructions (GEMINI.md, AGENTS.md, direct prompts) take precedence over skills. Only bypass skill workflows or Ponytail rules when explicitly instructed by the user.
04 同仓库更多 Skills
web-artifacts-builder
Suite of tools for creating elaborate, multi-component claude.ai HTML artifacts using modern frontend web technologies (React, Tailwind CSS, shadcn/ui). Use for complex artifacts requiring state management, routing, or shadcn/ui components - not for simple single-file HTML/JSX artifacts.
task-observer
Monitors task execution for skill improvement opportunities: logs patterns, user corrections and methodology worth preserving to the workspace, turns them into skill updates in a weekly review, and flags repeating work as new skills. Use during ANY multi-step task, agentic workflow, or work session. Also triggers in post-task feedback discussions and when the user mentions skill observations, the observation log, or skill taxonomy. Also known as "One Skill to Rule Them All" — trigger on this phrase too. IMPORTANT: invoke this skill before the FIRST tool call of any session and before writing or proposing a plan — any turn that will involve a tool call counts. This sentence is the session-start trigger and the only activation layer that survives an unreachable config file; pair it with a CLAUDE.md instruction or a session-start hook (references/environments.md). A subagent dispatched by a session already running it does not run it: it writes nothing and puts its findings in its report.
stress-tested-architecture
Combines 'brainstorming' and 'grilling'. Interactive design, RFC, and system architecture workflow that couples collaborative exploration with adversarial stress-testing. Features Three-Path Classification (Spike, Bounded, Architectural), a branching Design Tree, structured Grilling Rounds (The Frontier) with recommended answers, and rigorous scrutiny on failure modes, scale, and trade-offs before locking specifications. Triggers on /stress-tested-architecture, /grill-architecture, "design architecture", "brainstorm and grill", "stress test plan", or "system design RFC".
skill-creator
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
precision-diagnostics
Combines 'systematic-debugging' and 'diagnosing-bugs'. Strict multi-phase root-cause diagnosis protocol for hard bugs, intermittent failures, test breakages, and performance regressions. Enforces building a tight deterministic feedback loop (repro), multi-boundary instrumentation, ranked falsifiable hypotheses, and minimal root-cause fixes with regression tests before any code edits. Triggers on /precision-diagnostics, /diagnose, "diagnose bug", "debug hard bug", "systematic debug", or any persistent/intermittent software failure.
pptx-generator
Generate, edit, and read PowerPoint presentations. Create from scratch with PptxGenJS (cover, TOC, content, section divider, summary slides), edit existing PPTX via XML workflows, or extract text with markitdown. Triggers: PPT, PPTX, PowerPoint, presentation, slide, deck, slides.