agentleFS
Memories-ai-labsMemories-ai-labs/Luci-skills/skills/official/luci-profile/SKILL.md

luci-profile

Skill8 starsChanged 4 months agoMITView on GitHub

Write or rewrite ~/.Life/profile.md, the user's Luci profile, from their daily reports, meeting write-ups and entities. Use when a distill run says the profile is due, or when the user asks to write, refresh or redo their Luci profile.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    First profile: read every daily file, meeting write-up and entity in the folder.

    line 24
  2. Distinctive:

    If the prompt gives today's date, the user's name, or questions the user already answered, use them as given.

    line 29
  3. Very specific:

    Before you overwrite an existing profile.md, keep a copy: if ~/.Life/profile/history/<its updated date>.md does not exist, copy the current …

    line 34
  4. Very specific:

    Follow ~/.Life/RULES.md: anything it says to leave out stays out of the profile too, and anything under "Answers you gave Luci" is settled …

    line 45
  5. Very specific:

    If the prompt names the user, keep the name: line it gives in the frontmatter exactly as written.

    line 48
  6. Very specific:

    If the current profile.md has a ## Confirmed by you section, the user confirmed those lines: treat them as facts, and copy the section …

    line 50
  7. Very specific:

    Every Working on item carries 1-3 ^[YYYY-MM-DD HH:MM] moments taken from the daily files' blocks.

    line 111

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md
<!-- luci-skill-version: 1.0.0 -->

Luci profile

Luci's Life page reads ~/.Life/profile.md as data: the frontmatter keys are fields, the ## headings are fixed, and each bullet is split into parts on its first —, its · and its [ / ]. Keep that shape exactly, or the page drops the line.

Windows: %USERPROFILE%\.Life\profile.md.

When it runs

Usually at the end of a distill, after the daily files and entities are written. The distill prompt then says which kind of write this is:

  • First profile: read every daily file, meeting write-up and entity in the folder.
  • Weekly rewrite: the same, weighting the last 7 days most, and add the ## This week, you changed section at the end.

If the prompt gives today's date, the user's name, or questions the user already answered, use them as given. Asked directly with no distill prompt, write a first profile unless one already exists, in which case do a weekly rewrite.

Before you overwrite an existing profile.md, keep a copy: if ~/.Life/profile/history/<its updated date>.md does not exist, copy the current file there (create the folders if needed). The Life page reads that copy to show what changed. When Luci started the run it has usually made the copy already; then leave it alone. Then overwrite profile.md.

Rules

  • It describes the user to their own coding agent and to them. Second person, present tense, objective, and only what the files support. Never invent a number, a name or a project.
  • Follow ~/.Life/RULES.md: anything it says to leave out stays out of the profile too, and anything under "Answers you gave Luci" is settled — use it and never ask it again. Never write to that file; Luci adds the answers.
  • If the prompt names the user, keep the name: line it gives in the frontmatter exactly as written. Otherwise leave name: out.
  • If the current profile.md has a ## Confirmed by you section, the user confirmed those lines: treat them as facts, and copy the section unchanged right after Not sure.

Shape

---
luci_profile: 1
name: <only when the prompt gives it>
updated: <today, YYYY-MM-DD>
days: <number of daily files you read>
hours: <recorded hours across them, one decimal>
make: <0-100: share of working time spent making things (code, writing, design) rather than meeting, chatting, or reviewing>
switch: <0-100: how often focus switches between apps and tasks; 0 = long deep blocks, 100 = constant switching>
peak_hour: <0-23: local hour that starts the most productive window>
headline: >
  <one sentence about what they do and how, under 30 words>
---

## Summary

<3-4 sentences. Wrap the 2-3 phrases a reader may want proof for in **bold**.>

## Working on

- [<Project name>](<project-entity-slug>) · <hours>h — <what it covers> ^[YYYY-MM-DD HH:MM] ^[YYYY-MM-DD HH:MM]

## How you work

- <short figure, e.g. 27 min> — <one sentence: focus blocks, peak window, share of chat/social apps, meetings>

## People

- [<Name>](<person-entity-slug>) — <what they are to the user, from the notes>

## Tools

- <App or tool name>

## Not sure

- <A yes/no question about something you inferred but cannot confirm>? — <why you think so> [<Answer if your guess is right> / <Answer if it is wrong>]

Weekly rewrite only, at the end:

## This week, you changed

<One sentence comparing this week with the previous profile.>
- + <new project or person> — <detail>
- ↑ <something that grew> — <before → after>
- ↓ <something that shrank> — <before → after>
- − <something that went quiet> — <detail>

Limits

  • Up to 4 Working on items, 4 How you work lines, 5 people, 8 tools and 2 Not sure questions.
  • Every Working on item carries 1-3 ^[YYYY-MM-DD HH:MM] moments taken from the daily files' blocks.
  • In each Not sure line the first answer means your guess is right and the second means it is wrong.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.