agentleFS
Sign inStart free
hasuwini77hasuwini77/agentille/skills/agentille-init/SKILL.md

agentille-init

Skill1 starsChanged 20 days agoMITView on GitHub

Optional setup for agentille. Asks up to 5 skippable questions (name/role, delivery style, tone, when to ask first, never-do list) and merges the answers into ~/.agentille/profile.json so agents sound like you. /agt works without it. Runs only when the user types /agentille-init; model invocation is disabled.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Read ~/.agentille/profile.json (cat ...

    line 13
  2. Distinctive:

    Ask the questions below in one message.

    line 14
  3. Very specific:

    Merge, never replace: Read-merge-write: load the profile, set only the fields the user answered, and write it back (2-space indent, …

    line 20

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

agentille-init — optional voice profile

/agt runs fine with no profile at all. This skill only makes agents sound like you. It runs when the user types /agentille-init; never auto-trigger.

Steps

  1. Read ~/.agentille/profile.json (cat ... 2>/dev/null). Absent → start from {}. Present but not valid JSON → tell the user to fix or delete it, and STOP; never overwrite a corrupt file.
  2. Ask the questions below in one message. Each shows its current value (or the default) in brackets; an empty reply or "skip" keeps it.
    1. name / role — who are you, in a few words? [current or skip]
    2. deliveryStyle — direct | detailed | step-by-step | short-paragraphs [direct]
    3. tone — peer-to-peer | mentor | formal | blunt | casual [peer-to-peer]
    4. preTaskQuestioning — never | ambiguous-only | always [ambiguous-only]
    5. neverDo — things agents must never do, comma-separated [none]
  3. Merge, never replace. Read-merge-write: load the profile, set only the fields the user answered, and write it back (2-space indent, trailing newline). Keep every other key untouched: unknown keys, projects, routing, display, thinkingDepth. Never write the file wholesale from scratch. Skipped questions write nothing.
  4. Print the path and the resulting values in one short block. Re-running asks again with the saved values as defaults.

Profile fields

FieldMeaningDefault
name, rolewho the agents work forunset
deliveryStyledirect / detailed / step-by-step / short-paragraphsdirect
tonepeer-to-peer / mentor / formal / blunt / casualpeer-to-peer
preTaskQuestioningnever / ambiguous-only / alwaysambiguous-only
neverDolist of hard prohibitions, passed to every agent[]
thinkingDepthalways / complex-only / quick; edit by handcomplex-only
routing{autoFable, maxFablePerRun, fableWeeklyCeiling, maxRunUsd, stopAtCeiling, stopAtWeeklyPct, advisorOnOpus}; edit by handbuilt-in
display{deck: auto|off, focus: true|false, highlight: on|all|off, band: on|off, statusLine: on|off, mascot: on|off, rows: on|off, spinner: on|off}; edit by hand{deck: off, focus: true, highlight: off, band: on, statusLine: on, mascot: on, rows: on, spinner: on}
sessionson / off: off stops a plain prompt from opening sessions (a typed /storm still can); edit by handon
viewportsdesign-reviewer default, any of desktop / tablet / mobile; edit by handdesktop + mobile
projects[]repos registered by /agentille-project[]

Old profiles keep working: fields nobody reads are ignored, and there are no migrations.

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.

Check your setup, free