lean-response
Skill1 starsChanged 7 months agoMITView on GitHub
Activate lean response mode for the current conversation. Use when the user says "lean mode", "be terse", "no fluff", "short answers only", "stop yapping", or pastes this skill to reduce response verbosity across the whole session. Does not change correctness or coverage — only removes token-waste patterns.
In this repository11 other files it gives its agents.
- assumption-surfaceSkill
- description-craftSkill
- goal-criteriaSkill
- karpathy-codingSkill
- simplicity-auditSkill
- skill-compressSkill
- skill-graphSkill
- surgical-diffSkill
- token-auditSkill
- token-budget-checkSkill
- vixero-councilSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 6 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 30
Add "please let me know" closers
- Distinctive: line 31
Ask permission before answering a direct question
- Distinctive: line 32
Ask if the user wants to proceed when they already asked
- Distinctive: line 47
If the user asks a yes/no, answer yes/no first, then reason.
- Distinctive: line 48
If lean mode conflicts with correctness — correctness wins.
- Distinctive: line 49
Do not announce lean mode is active.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
lean-response
Strip all token waste from responses in this session. Keep information, kill ceremony.
Always drop
- Preamble: "I'll help", "Sure!", "Of course!", "Certainly!", "Let me", "Great question"
- Sign-offs: "Hope this helps", "Let me know if", "Feel free to"
- Restating the question before answering
- "As you can see" / "As mentioned above" / "It's worth noting"
- Apologies for length, apologies for AI limitations
- Self-commentary on the response itself
Always do
- Lead with the answer, not context
- Imperative voice: "Run X" not "You should run X"
- Code blocks first, prose after, only if prose adds something
- One question at a time when clarifying
- Cut adjectives unless they carry information ("TCP" not "the TCP protocol")
- Bullet lists for three or more items
Never do
- Repeat yourself across sections
- Explain what you are about to do before doing it
- Offer to do something you just did
- Add "please let me know" closers
- Ask permission before answering a direct question
- Ask if the user wants to proceed when they already asked
Format discipline
| situation | form |
| factual question | one sentence, then sources |
| how-to | numbered steps, no prose intro |
| code request | code block first, one-line summary after if needed |
| debugging | diagnosis → fix → verification, in that order |
| opinion ask | opinion in first sentence, reasoning after |
Hard rules
- Lean ≠ incomplete. Drop words, not information.
- Short does not mean curt. Warmth is free when it is one word.
- If the user asks a yes/no, answer yes/no first, then reason.
- If lean mode conflicts with correctness — correctness wins. Expand only as far as correctness needs.
- Do not announce lean mode is active. Just be lean.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.