mjdev-user-requests-first
Forces the assistant to prioritize user requests above its own TODO lists or plans.
Install
mkdir -p .claude/skills/mjdev-user-requests-first && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/18172" && unzip -o skill.zip -d .claude/skills/mjdev-user-requests-first && rm skill.zipInstalls to .claude/skills/mjdev-user-requests-first
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
User requests always take priority over the assistant's own plans, task lists, or agenda. Never ignore, defer, or override what the user asked in favor of self-generated work. When the request is unclear or ambiguous, ASK a question instead of guessing or doing your own thing. Apply on every turn in this project.Key capabilities
- →Prioritize user's stated request over assistant's own plans
- →Execute user requests in the order they were asked
- →Ask one focused question when user request is unclear or ambiguous
- →Finish the current requested item before opening new fronts
- →Keep the user's workspace clean by not committing self-generated changes without explicit approval
How it works
The skill enforces a strict rule that user requests take precedence over any self-generated work, requiring direct execution or a clarifying question if ambiguous.
Inputs & outputs
When to use mjdev-user-requests-first
- →Preventing unwanted task execution
- →Refocusing on specific user instructions
- →Stopping auto-generated agendas
About this skill
User requests come first (non-negotiable)
The user's stated request is the priority. My own TODO list, "recommended order", plans, or ideas are subordinate to what the user actually asked for right now.
Rules
- Do what the user asked, in the order they asked. If the user says "focus on X / don't touch Y / save it for later / stop changing code", that overrides any plan I had. Drop my agenda and serve the request.
- Never silently ignore or reinterpret a request. If the user says "test it yourself", I test it myself — I don't push it back on them. If they say "don't change anything", I change nothing.
- When unclear or ambiguous, ASK one focused question rather than guessing, assuming, or doing something adjacent. A short question beats unwanted work.
- Don't pile on work. The user is frustrated when tasks accumulate instead of finishing. Finish the current requested item cleanly before opening new fronts. Don't task-manage the user.
- Keep their workspace clean. Don't commit / leave my own changes (esp. build-config or scratch) in git without an explicit greenlight.
Why
The user has repeatedly had to repeat themselves ("otestuj si to", "neměň nic", "řeším gif") because self-directed work drifted from their request. That wastes their limited patience and rebuild time. Their request is the spec; my plans are only suggestions until they accept them.
See [[autonomy-only-ask-when-unsure]] and [[feedback-test-it-yourself]].
When not to use it
- →When the user explicitly defers a request
- →When the user explicitly asks the assistant to follow its own plan
- →When the user wants the assistant to guess or do something adjacent to an unclear request
Limitations
- →The skill does not allow the assistant to silently ignore or reinterpret a request
- →The skill prohibits piling on work by opening new fronts before finishing current items
- →The skill prevents committing self-generated changes without explicit user approval
How it compares
This skill explicitly overrides the assistant's internal agenda and task lists in favor of immediate user requests, ensuring direct adherence to user instructions, unlike an assistant that might prioritize its own workflow.
Compared to similar skills
mjdev-user-requests-first side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| mjdev-user-requests-first (this skill) | 0 | 1mo | No flags | Beginner |
| sequential-thinking | 136 | 8mo | No flags | Intermediate |
| ai-wrapper-product | 5 | 6mo | No flags | Intermediate |
| behavioral-modes | 3 | 6mo | No flags | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
sequential-thinking
mrgoonie
Use when complex problems require systematic step-by-step reasoning with ability to revise thoughts, branch into alternative approaches, or dynamically adjust scope. Ideal for multi-stage analysis, design planning, problem decomposition, or tasks with initially unclear scope.
ai-wrapper-product
davila7
Expert in building products that wrap AI APIs (OpenAI, Anthropic, etc.) into focused tools people will pay for. Not just 'ChatGPT but different' - products that solve specific problems with AI. Covers prompt engineering for products, cost management, rate limiting, and building defensible AI businesses. Use when: AI wrapper, GPT product, AI tool, wrap AI, AI SaaS.
behavioral-modes
davila7
AI operational modes (brainstorm, implement, debug, review, teach, ship, orchestrate). Use to adapt behavior based on task type.
token-budget
toonight
Manages token budget estimation and tracking to prevent context overflow
adversarial-spec
zscole
Iteratively refine a product spec by debating with multiple LLMs (GPT, Gemini, Grok, etc.) until all models agree. Use when user wants to write or refine a specification document using adversarial development.
initializing-memory
letta-ai
Comprehensive guide for initializing or reorganizing agent memory. Load this skill when running /init, when the user asks you to set up your memory, or when you need guidance on creating effective memory blocks.