agentleFS
localoy-ailocaloy-ai/localstack/social-post/SKILL.md

social-post

Skill2 starsChanged 41 days agoMITView on GitHub

Writes the caption for each post in the social calendar, says when a post needs a picture from Nova (/vinci), and posts each one only after the person says yes to it, through their own signed-in browser — never signing in. (localstack)

19 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

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

  • 0

    Common

    Most of the 15 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    The right account: Before the yes, check the account name the platform shows matches the plan's account.

    line 51
  2. Distinctive:

    Every posted entry carries its live URL.

    line 148

“Common” is measured: kept by at least 6 in 10 of the 15 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

When to invoke this skill

Writes each planned post's caption into the topic's PLAN-<topic>.md, and then — only when the person asks — posts them one at a time from their own browser. Use after /social-plan, or when asked to "write the captions", "post this", "publish this week's posts". Fizz's skill.

One yes per post — the hard boundary

  • Every post waits for its own yes. Show the channel, the account, the exact caption and the picture file, then ask. "Post them all", "yes to the rest" and silence are not a yes for the next one. Auto mode skips ordinary questions, never this one.
  • The person's own session, never a new one. The browser is already signed in to that account, or it is not. Never type a password, create an account or solve a CAPTCHA. A login wall ends posting for that channel, logged as blocked.
  • The right account. Before the yes, check the account name the platform shows matches the plan's account. A mismatch stops that channel.
  • A platform warning is a stop. A rate limit, spam notice or restriction banner ends the run, logged and told to the person.
  • No browser, no posting. Say so and leave the captions written. Never fall back to an API, a scheduler tool or a script.

Pictures come from Nova

A row marked needs picture gets a caption, not an invented image. Say so in plain words — "this post needs a picture: ask Nova with /vinci <brief>" — with a one-line brief (what, size for the channel, words on it), and add it as a TODO. A post whose picture is still missing is not offered for posting. A picture the person supplied (have: <file>) is used as given.

What you read first

  • The plan: ls PLAN-social*.md 2>/dev/null — if the person named one, use it; one → use it; several → ask which; none → offer /social-plan first (or take a single post the person dictates, logged in CHANGELOG.md only).
  • The calendar rows with Status: planned (to write) or written (to post).
  • DESIGN.md — voice, hashtags policy, words to avoid. It wins over your defaults.
  • Each row's source — the caption's facts come from it, nowhere else.

Ground rules (non-negotiable)

  1. Facts from the source. A number, a quote, a claim about a customer or result appears only if the row's source shows it.
  2. Fit the channel. Length and format per the channel's limits; hashtags only as DESIGN.md allows.
  3. Posted exactly as approved. The text posted is the caption shown at the yes, word for word.

Procedure

1. Load the inputs above.

Scratch for this run lives in .localstack/work/{date}-{slug}/ — hidden, one directory per run, so a new run never clobbers an earlier one and the folder's top level stays the standard files. Scratch is disposable; old run directories may be deleted freely.

2. Write captions. Under ## Posts in the plan (create it after the calendar; never remove an entry), per row:

### <Day> — <Channel> — <idea>
Status: written | waiting for picture
- Source: <URL/file>
- Picture: <file, or "needs picture — /vinci <brief>">

<caption>

Set the calendar row's Status to written.

3. Post — only when asked, one at a time. For each written post with its picture ready: open the channel's composer in the signed-in browser, attach the picture, put the caption in exactly, check the account name, and stop:

<Channel> — <account>
<the exact caption> · picture: <file>
Post this?

On yes, post, then confirm it is live (the post on the profile or a posted notice) and note its URL. Edit → show again and ask. Skip → skipped. Status becomes posted YYYY-MM-DD <post URL>, skipped, failed — <what you saw> or blocked — <why>, in ## Posts and the calendar row.

Standard files. This folder is kept in files any agent already reads. Update them in place; never scatter output into new folders.

  • AGENTS.md — create it if missing. localstack owns only the block between <!-- localstack:start --> and <!-- localstack:end -->; rewrite that block, never anything outside it. The block says what this folder is for, the rules (drafts only; nothing is sent without the user's explicit yes, one message at a time; no invented facts), a map of the files below, and one line per topic (its PLAN, its lead count, the next unticked step) and per report (its file and date).
  • PLAN-<topic>.md — in ## Steps, tick - [x] post (add today's date after it) once this step is done; leave it unticked if you stopped partway, and say why in CHANGELOG.md.
  • CHANGELOG.md — create it if missing (# Changelog). Add one bullet for this run under today's ## YYYY-MM-DD heading, newest date first: the skill, the topic, and the counts or outcome (e.g. - lead-search austin-dentists: 18 found, 3 skipped as already contacted).
  • TODOS.md — create it if missing (# TODOs). Add each open next action as - [ ] <action> (<topic>); tick items this run finished; never delete lines.
  • DESIGN.md — decisions meant to last (positioning, tone, channels to use or avoid). Read it before writing anything a person will see; add to it only when the user states or approves a decision.

For this step: tick - [x] write captions when every row is written; the standard tick above marks post only when no post is left unposted. The CHANGELOG line gives written / posted / skipped / blocked counts, and each posted caption with its URL. TODOs: one per missing picture, one per blocked channel.

4. Report. What was written, what was posted (with links), what waits for a picture or a yes.

Completion status. End the chat report with one of:

  • DONE — completed, with the evidence named (files written, counts, URLs).
  • DONE_WITH_CONCERNS — completed, and list each concern.
  • BLOCKED — cannot proceed; say what blocked it and what was tried.
  • NEEDS_CONTEXT — missing information; say exactly what is needed.

Quality bar

  • Zero posts without their own yes, zero sign-ins, zero invented claims.
  • Every posted entry carries its live URL.
  • A missing picture is a named TODO for Nova, never a placeholder posted.

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.