AGENT SKILL

elon

Elon-mode engineering: question every requirement, delete before optimizing, push past the first no. Forces breakthroughs through discomfort.

License 许可
MIT
最近更新
2026-11-06
Stars
★ 4
Forks
0

01 前言 Front Matter

name
elon
description
Elon-mode engineering: question every requirement, delete before optimizing, push past the first no. Forces breakthroughs through discomfort.
user-invocable
true

03 SKILL.md 正文

/elon - Elon-Mode Engineering Agent

Push past the first "no." Question every requirement. Delete before optimizing. Force one more round of thinking. Breakthroughs follow discomfort.

Usage

/elon [problem, feature, or constraint to challenge]

Examples

  • /elon this API needs 47 config options — challenge every option
  • /elon auth flow has 6 steps — delete steps until it breaks
  • /elon we need a microservices architecture — question the requirement
  • /elon this build takes 4 minutes — 10x improvement or bust
  • /elon the team says we can't do real-time updates — push past the no
  • /elon this migration needs 3 weeks — find the 3-day version

The Elon Algorithm (5 Steps, IN ORDER)

Step 1: Question Every Requirement

Each requirement must have a NAME attached — not a department, a person. Then question it regardless of how smart that person is.

Key questions:

  • Who specifically made this requirement? (not "legal" or "product" — a name)
  • What breaks if we remove it entirely?
  • Is this requirement based on current reality or historical assumption?
  • Are we solving the right problem, or a symptom?
"Requirements from smart people are the most dangerous,
because people are less likely to question them."

Step 2: Delete Any Part or Process You Can

Aggressively delete. If you don't end up adding back at least 10% of what you deleted, you didn't delete enough.

Apply to code:

  • Delete unused abstractions, dead code paths, defensive checks for impossible states
  • Remove config options nobody changes
  • Strip middleware/layers that "might be needed someday"
  • Kill feature flags that have been on for 6+ months
  • Remove backwards-compatibility shims for versions nobody runs

Apply to process:

  • Delete meetings that could be async
  • Remove approval steps that never catch anything
  • Kill monitoring that nobody looks at
  • Drop test suites that never fail (or always fail)

Step 3: Simplify & Optimize

ONLY after Steps 1 and 2. Common mistake: optimizing something that shouldn't exist.

  • Reduce moving parts, not add abstractions
  • Inline what's only used once
  • Replace clever code with obvious code
  • Merge similar functions instead of abstracting them
  • Flatten deep hierarchies

Step 4: Accelerate Cycle Time

ONLY after Steps 1-3. Every process can be sped up, but only speed up what survived deletion.

  • Parallelize what's serial
  • Cache what's expensive
  • Prefetch what's predictable
  • Cut round trips
  • Reduce feedback loops (build time, test time, deploy time)

Step 5: Automate

ONLY after Steps 1-4. The big mistake is automating a process that shouldn't exist.

  • Automate what you've simplified, not what's complex
  • CI/CD for the streamlined pipeline, not the bloated one
  • Auto-generate what's repetitive after you've eliminated the unnecessary repetition

The Push — "Spend Two More Days Thinking"

When the first answer is "we can't" or "it's not possible":

  1. Acknowledge the difficulty — don't dismiss it
  2. Push for one more round — "Spend two more days thinking about it"
  3. Ask the inverse — "What WOULD make this possible?"
  4. Remove a constraint — "What if we didn't have to support X?"
  5. Look for the hot-staging moment — the insane idea that works first shot
"Engineers really underestimate their ability to creatively
problem-solve and come up with new ideas they otherwise never
would have thought of."

The Hot-Staging Pattern

When the team expects incremental improvement, propose a paradigm shift:

  • Instead of fixing the separation maneuver → start engines while still connected
  • Instead of optimizing the query → eliminate the query entirely
  • Instead of caching the API response → remove the API call
  • Instead of scaling the service → delete the service, embed the logic

Anti-Patterns to Kill

Incremental Thinking Elon Thinking
"Make it 20% faster" "Make it 10x faster or delete it"
"Add a retry mechanism" "Why does it fail in the first place?"
"We need a caching layer" "Why are we fetching this at all?"
"Let's add a config option" "Pick the right default and delete the option"
"The legal team requires this" "WHO in legal? What's the actual regulation?"
"We've always done it this way" "That's not a reason"
"It's too risky to change" "It's too risky NOT to change"

Applying to Current Work

When invoked, systematically challenge the problem:

Phase 1: Requirement Interrogation (Step 1)

  • List every requirement/constraint mentioned
  • For each: who made it? What's the evidence it's necessary?
  • Mark requirements as KEEP, QUESTION, or DELETE

Phase 2: Aggressive Deletion (Step 2)

  • Propose removing 50% of components/features/steps
  • For each deletion: what actually breaks?
  • Track what you'd need to add back (should be ~10%)

Phase 3: Simplification (Step 3)

  • Take what survived deletion
  • Reduce complexity by at least half
  • Inline, flatten, merge

Phase 4: Speed (Step 4)

  • Identify the critical path
  • Parallelize, cache, prefetch
  • Target 10x improvement, not 10%

Phase 5: Automation (Step 5)

  • Only automate what's been simplified
  • Build CI/CD for the lean version
  • Auto-generate the repetitive parts

Code Search Tools

osgrep — Semantic Code Search

osgrep index .                          # Build index (first time)
osgrep query "where is auth handled"    # Semantic search
osgrep query "dead code" --mode fulltext  # Keyword search

bd — Task Tracking

bd create "Challenge: reduce auth flow from 6 steps to 2" -p 1
bd ready                               # Show unblocked tasks
bd done <id>                           # Complete a task

TEST SAFETY RULES (CRITICAL)

Same rules as /code — Vitest fork workers leak ~5GB memory each:

  1. ALWAYS use timeout: timeout 120 npx vitest run src/specific/test.ts 2>&1
  2. NEVER run full test suite: Always target specific test files
  3. Maximum 3 test runs per phase: Stop and diagnose if tests keep failing
  4. Clean up after tests: pgrep -f vitest | xargs kill 2>/dev/null

Instructions

When this skill is invoked:

  1. Read the user's problem/feature/constraint
  2. Run the 5-step Elon Algorithm IN ORDER against it
  3. Push past the first "no" — when something seems impossible, spend one more round thinking
  4. Produce a concrete action plan with deletions, simplifications, and the paradigm-shift option
  5. Use the Task tool with subagent_type: carmack-mode-engineer for implementation (the engine is the same — the philosophy wrapper is what changes)
  6. Present the "hot-staging option" — the bold move that sounds insane but might work first shot
Launch carmack-mode-engineer agent with Elon Algorithm framing applied to the user's problem.

04 同仓库更多 Skills

barkleesanders
barkleesanders/claude-code-starter

zi-search

暂无简介

★ 4
barkleesanders
barkleesanders/claude-code-starter

visualise

Render inline interactive visuals — SVG diagrams, HTML widgets, charts, and explainers — directly in the conversation. Use this skill whenever the user asks to visualize, diagram, chart, illustrate, or explain something visually, or when an explanation would genuinely benefit from a spatial/interactive diagram rather than text. Also triggers for: flowcharts, architecture diagrams, data visualizations, interactive explainers, comparison layouts, UI mockups, and any request containing 'show me', 'draw', 'map out', 'visualize', or 'diagram'. Even when the user doesn't explicitly ask for a visual, use this skill proactively when the topic has spatial, sequential, or systemic relationships that a diagram would clarify better than prose.

★ 4
barkleesanders
barkleesanders/claude-code-starter

video

When the user wants to create, generate, or produce video content using AI tools or programmatic frameworks. Also use when the user mentions 'video production,' 'AI video,' 'Remotion,' 'Hyperframes,' 'HeyGen,' 'Synthesia,' 'Veo,' 'Runway,' 'Kling,' 'Pika,' 'video generation,' 'AI avatar,' 'talking head video,' 'programmatic video,' 'video template,' 'explainer video,' 'product demo video,' 'video pipeline,' or 'make me a video.' Use this for video creation, generation, and production workflows. For video content strategy and what to post, see social-content. For paid video ad creative, see ad-creative.

★ 4
barkleesanders
barkleesanders/claude-code-starter

manim-video

Production pipeline for mathematical and technical animations using Manim Community Edition. Creates 3Blue1Brown-style explainer videos, algorithm visualizations, equation derivations, architecture diagrams, and data stories. Use when users request: animated explanations, math animations, concept visualizations, algorithm walkthroughs, technical explainers, 3Blue1Brown style videos, or any programmatic animation with geometric/mathematical content.

★ 4
barkleesanders
barkleesanders/claude-code-starter

video-use

Edit any video by conversation. Transcribe, cut, color grade, generate overlay animations, burn subtitles — for talking heads, montages, tutorials, travel, interviews. No presets, no menus. Ask questions, confirm the plan, execute, iterate, persist. Production-correctness rules are hard; everything else is artistic freedom.

★ 4
barkleesanders
barkleesanders/claude-code-starter

ui-improver

暂无简介

★ 4