aim — Claude Code skill: reads data, proposes the intent
A Claude skill for the moment files, a repo or a chat scrollback arrive with a thin ask. It reads everything and proposes what you meant, so you only confirm.
npx skills add LilMGenius/paperthin --skill aim -a claude-codeWhat aim does
aim changes what the agent does when the material has already landed but the request is vague. Instead of asking a clarifying question, it reads the handed-over data end to end and states the intent it reads in one line, grounded in what the data actually shows. Then it says briefly what it would do next.
The workflow is four steps: read everything first, name the intent, name the next move, ask only to confirm. The turn ends with a proposal you can accept, adjust or reject. If the data is too thin to propose from, the skill says so and stops rather than falling back to a question.
Everything lives in a single SKILL.md with the goal, the workflow, the rules and a verification section. The check is blunt: if your next reply has you explaining what you wanted from scratch, the proposal was a question in disguise. One intent per turn — when two are plausible, the likelier one is proposed and the other gets a clause.
What it helps with
- figure out what to do with a folder of files dropped without context
- read intent from a repo shared as a bare link
- recover the task from a long chat scrollback
- get a ready proposal instead of a clarifying question
- skip composing a request from scratch
Example prompts for the agent
- here's the repo, take a look and tell me what's in it for me
- I dropped logs and a couple of screenshots, make sense of it
- read this thread and propose what I'm after
How to install aim
1Skills CLI (Vercel)
npx skills add LilMGenius/paperthin --skill aim -a claude-code2Manually: put the files in place
~/.claude/skills/aim/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/aim/
1Skills CLI (Vercel)
npx skills add LilMGenius/paperthin --skill aim -a codex2Manually: put the files in place
.agents/skills/aim/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add LilMGenius/paperthin --skill aim -a cursor2Manually: put the files in place
.cursor/skills/aim/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install LilMGenius/paperthin aim2Manually: put the files in place
.github/skills/aim/SKILL.md1Skills CLI (Vercel)
npx skills add LilMGenius/paperthin --skill aim -a gemini-cli2Platform CLI
gemini skills install https://github.com/LilMGenius/paperthin --path skills/depth/aimManually: put the files in place
.devin/skills/aim/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add LilMGenius/paperthin --skill aim -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about aim
How do I install aim in Claude?
Run npx skills add LilMGenius/paperthin --skill aim -a claude-code. The skill file lands at ~/.claude/skills/aim/SKILL.md.
What does aim do?
It reads the handed-over data in full and proposes the intent behind it in one line, pointing at what it read. Then it names the next step and asks only for confirmation, a correction or an adjustment.
How is aim different from a clarifying question?
A question makes you compose a sentence you did not have ready. aim hands you a proposal so you can answer with a yes or a small edit.
Source text
SKILL.md · MIT
--- name: aim description: "When data lands with a thin ask — files, a repo, a paste, a chat scrollback — read it and propose what the user is trying to do with it, as a proposal to confirm rather than a question to answer. Use the moment material arrives without a fully-formed ask, or when the ask reads thinner than the data suggests." --- Read the data and propose the intent behind it, for the user to confirm. ## Goal The result is a proposal the user only confirms, revises, or rejects — never a question. Asking makes the user compose a sentence they did not have ready; proposing lets them answer with a yes. If the turn ends with them writing that sentence anyway, the skill missed. ## Workflow 1. Read every piece of the handed-over data end to end before proposing anything. 2. State the intent you read, in one line grounded in what the data actually shows. 3. State briefly what you would do next on that intent. 4. Ask only to confirm — right, wrong, or adjust. ## Rules - Read before proposing. A proposal without reading is a guess. - Ground it in the data. Point at what you read, not what you assumed. - One intent per turn. If two are plausible, propose the likelier and name the other in a clause. - Confirm, don't compose. If the turn ends with the user assembling a sentence, redo the proposal. - If the data is too thin to propose from, say so and stop. Do not fall back to asking. ## Verification The user's next reply is short — a yes, a correction, or a small adjustment. If they had to explain what they wanted from scratch, the proposal was wrong, or was really a question in disguise.
© LilMGenius · MIT · LilMGenius/paperthin
