agentleFS
Sign inStart free
vidit19sharmavidit19sharma/neogents/skills/train/SKILL.md

train

Skill1 starsChanged 11 days agoMITView on GitHub

Promote recurring workflows from the second brain into project skills. Reviews .neo/brain/WORKFLOWS.md candidates, drafts a SKILL.md for the chosen workflow, and writes it to the project's .claude/skills/ after user approval. Use when neo-shadow flags a repeated workflow or the user says "make this a skill" or "train that workflow".

8 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 1

    Distinctive

    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
  1. Very specific:

    Guard: if .neo/brain/WORKFLOWS.md is missing or has no entries, stop — nothing to train on yet.

    line 12
  2. Very specific:

    Review the ledger: read .neo/brain/WORKFLOWS.md.

    line 14
  3. Distinctive:

    Keep it under a screen. A skill that needs scrolling needs splitting.

    line 20

“Common” is measured: kept by at least 6 in 10 of the 8 edited versions of this file we could read. “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

/neo:train — turn repeated workflows into skills

Candidates: !if test -f .neo/brain/WORKFLOWS.md && test ! -L .neo/brain/WORKFLOWS.md; then head -c 1048576 .neo/brain/WORKFLOWS.md | grep -c '^## candidate: [^<]' || true; else echo "0 (no WORKFLOWS.md — run /neo:init or wait for neo-shadow to record some)"; fi · Existing project skills: !ls -q .claude/skills 2>/dev/null || echo "none"

Steps

  1. Guard: if .neo/brain/WORKFLOWS.md is missing or has no entries, stop — nothing to train on yet. Tell the user neo-shadow records candidates automatically at save time.
  2. Review the ledger: read .neo/brain/WORKFLOWS.md. List entries to the user, proposed first (3+ sightings), then candidate. Skip skilled — already promoted. If the user named a workflow, go straight to it; a user request overrides the 3-sighting threshold.
  3. Draft the skill for the chosen workflow:
    • Directory name = the candidate's kebab-name (that becomes the command name).
    • Frontmatter: name matching the directory, description built from the entry's trigger line — concrete enough that auto-invocation fires on the right requests.
    • Body: the steps from the ledger, generalized (strip session-specific paths/values into placeholders), each step with its verify command where one exists.
    • Keep it under a screen. A skill that needs scrolling needs splitting.
  4. Approval gate: show the user the full draft and the target path .claude/skills/<name>/SKILL.md (the project's skills directory — never the plugin's). Do not write without a yes.
  5. Write + record: on approval, first stop and tell the user if .claude, .claude/skills, .claude/skills/<name> or SKILL.md is a symlink ([ -L ... ]) — a clone can ship one pointing outside the repo, and the write would follow it. Then write the file yourself (you, NEO, on the main thread — this is outside the shadow's jail, like /neo:init writing BRIEF.md). Then spawn neo-shadow with a one-line delta: mark the entry status: skilled in WORKFLOWS.md and add a pointer to the new skill path.
  6. Report: "Trained: /<name> — invoke it directly or let auto-delegation pick it up. Restart or /reload-plugins not needed for project skills."

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