agentleFS
pradeep-sankaran-07pradeep-sankaran-07/team-os-gdrive-template/.agents/skills/_template/SKILL.md

_template

Skill0 starsChanged 2 months agoMITView on GitHub

One sentence on what this skill produces. Use when <the situation>. Triggers: "<the phrase someone actually types>", "<another one>", "<a third>". NOT for <the adjacent case> (use <sibling-skill> instead).

17 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

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

  • 0

    Common

    Most of the 8 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the skill has a real judgment call, isolate it into its own step and give the selection criteria.

    line 47
  2. Distinctive:

    If the skill runs a script, inline the exact invocation with every flag.

    line 54
  3. Very specific:

    Every skill ends the same way: write the output to output_location, refresh whatever summary file it belongs to, and append one line to …

    line 58
  4. Distinctive:

    Add to it whenever something produced a wrong output, and say what went wrong rather than only what to do instead.

    line 84

“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

<Skill name>

One paragraph: what this produces and who reads it. No preamble.

Step 1 - Gather

Name the exact files to read, in order. Say what to do when one is missing: record it as missing, never infer it.

Step 2 - Decide

If the skill has a real judgment call, isolate it into its own step and give the selection criteria. A skill whose steps are all "write section N" does not need this step; one that picks an angle, a template or a route does.

Step 3 - Produce

The sections of the output, numbered, with one line each on what belongs in them. If the skill runs a script, inline the exact invocation with every flag.

Step 4 - Record

Every skill ends the same way: write the output to output_location, refresh whatever summary file it belongs to, and append one line to _ops/log.md.

Boundary: when this does NOT apply

Name the adjacent cases and the skill that owns each. Be concrete. This section and the description boundary clause say the same thing twice on purpose, because one is read by the assistant while routing and the other while working.

Non-negotiables

Three to five lines. The rules that make the output wrong if broken, not general good practice.

Review before delivering (required)

A short checklist for the human who finishes the work. Every customer-facing skill has one. The skill does about 70 percent; a person adds the judgment.

Things that have gone wrong before

A running list, newest last, in the house shape:

  • <the rule>. Why: <the reason>. (<date>, <what happened>)

This section is the reason a skill gets better instead of just getting longer. Add to it whenever something produced a wrong output, and say what went wrong rather than only what to do instead.

Folders

  • references/ - longer material loaded on demand. Point at it by name from the steps above; do not inline it here.
  • assets/ - templates, tokens, images. Keep it small; the whole folder is mirrored and, for some owners, changes here escalate rather than auto-promote.

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.