agentleFS
priyanshu-aryapriyanshu-arya/LinkedIn_Automation/.claude/skills/publish-substack/SKILL.md

publish-substack

Skill0 starsChanged 59 days agoMITView on GitHub

Use when the user wants to move an approved Substack article or Note toward publishing, get the copy-ready final content, or explicitly invokes /publish-substack. Substack has no Buffer channel and no safe official publishing API, so this never calls an external API — it hands off a copy-ready Substack-Ready note and only marks something published once the user explicitly confirms they posted it themselves. Never claims a publish happened without that confirmation.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

6 rules. About 0 are common, 3 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Ask the user directly (AskUserQuestion when interactive): has this actually been published on Substack yet?

    line 54
  2. Very specific:

    If not yet published: leave it in Substack-Ready/ at publish_status: ready_to_publish.

    line 66
  3. Distinctive:

    Never call any Substack API, official or unofficial — this skill is manual-handoff by design, not a missing feature.

    line 79
  4. Very specific:

    Never set publish_status: published or publish_confirmed_date without an explicit user confirmation this run.

    line 81
  5. Distinctive:

    Never invent or guess a publish date from context (e.g.

    line 83
  6. Very specific:

    Never process a note that isn't status: approved.

    line 85

≈ “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

Publish Substack (manual publish handoff — Substack Articles + Notes)

Substack's sibling to /schedule-approved//schedule-approved-x, adapted for the one real constraint that makes those skills' approach impossible here: Substack has no Buffer channel and no supported API for posting. Rather than reach for an unofficial, session-cookie-based API (ToS-risky, undocumented, could break silently), this skill does what the pipeline already does for image generation — hand off a finished, ready-to-paste deliverable and let the human do the actual publish action, then record that confirmation honestly rather than assuming it happened.

Arguments

  • note-id (optional) — hand off one specific approved item. If omitted, process every status: approved note with platform in {substack-article, substack-note} and no ready_id yet.

Process (per note)

1. Move to Substack-Ready/

Copy _Templates/Substack-Ready-Note.md into Substack-Ready/ as YYYY-MM-DD--kebab-slug.md. Fill in draft_id, platform, and:

  • For substack-article: title, and the full article body copied into ## Final Content verbatim (in-body images already generated should be noted as "insert image: <label>" at their marker positions, since this handoff step doesn't embed files). Call out the header/hero image separately in ## Publishing Notes — it goes in Substack's dedicated header-image slot in the editor, not inline in the body, so don't leave it looking like just another in-body marker.
  • For substack-note: final_text and the same text mirrored into ## Final Content.

publish_status: ready_to_publish. Copy category/format/hashtags/ visual_ids/sources from the source note; compute length from the final text/body. Set the source Draft/Article note's ready_id to this new note's id.

2. Present the copy-ready deliverable

Show the full final content exactly as it should be pasted into Substack's editor, plus a reminder of any images still needing manual insertion at their marked spots. State plainly: this is not published yet — nothing external has happened, this is the same "prompt not a picture" honesty /generate-visual already applies, just for the whole piece.

3. Ask for publish confirmation

Ask the user directly (AskUserQuestion when interactive): has this actually been published on Substack yet? Do not assume yes because time has passed or because the user moved on to another task.

  • If confirmed published: set publish_confirmed_date to today, publish_status: published, move the file from Substack-Ready/ to Published-Posts/. Update the source Draft/Article note's status unchanged (it stays approved — the historical record, same pattern as /schedule-approved's Draft Notes) but append {action: scheduled, date, note: "published manually via publish-substack"} — reusing the existing history action name since it means the same thing here: this note left the approval stage for real, external distribution.
  • If not yet published: leave it in Substack-Ready/ at publish_status: ready_to_publish. This is a normal, expected state — not an error — for however long it takes the user to actually paste it in.

Report back

Which notes moved to Substack-Ready/, which were confirmed published this run (with their publish_confirmed_date), and which are still waiting on the user to actually publish them.

Hard rules

  • Never call any Substack API, official or unofficial — this skill is manual-handoff by design, not a missing feature.
  • Never set publish_status: published or publish_confirmed_date without an explicit user confirmation this run.
  • Never invent or guess a publish date from context (e.g. "they probably posted it by now") — ask.
  • Never process a note that isn't status: approved.

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.