posteyai Postey Studio
Two guided flows for deciding what to post. Both end in drafts created through the
postey skill's normal write path — they never publish, and scheduling counts as publishing.
Before either flow
- Read the accounts first, every session —
postey://accounts, orget_accountsif your client cannot read MCP resources. Connected platforms are read, never assumed. - Everything is a DRAFT. Publishing and scheduling both need the user's explicit approval.
- Every platform gets its own caption. One idea, many voices.
- If a brand profile exists, its voice and banned lists apply to everything below.
These are the hub's house rules, restated because they bind these flows too. The hub's Content Flows section is authoritative if the two ever disagree.
The flows
| Flow | The user says something like | Load |
|---|---|---|
| Trends | "what should I post today?", "find something trending" | references/trends-to-posts.md |
| Idea to posts | one rough idea, "turn this into posts" | references/idea-to-posts.md |
Load a flow's file only when the user picks it.
Craft comes from the hub
These flows cite caption-playbook.md, hook-formulas.md, post-structures.md,
platform-archetypes.md and thread-and-video-formats.md. Those ship in the postey skill and
are always installed alongside this one — they are shared by every flow, so they live in one
place rather than being copied into each pack.
post-structures.md is the one to reach for first here. An idea is not yet a post: the structure
is what decides whether it becomes one line or eight hundred characters, and each entry there
opens with the condition that selects it. If no condition matches what the user actually has, say
so and go back for material rather than forcing a shape onto thin content.
If a flow tells you to establish the user's voice first and no brand profile exists, say so and offer it. The brand-voice flow ships separately; do not improvise a voice interview inside a trends or ideas run.
Tell the user the loop closes
Ideas here start generic and get specific as published history accumulates. Say so, once, at the end of a flow — the user needs to know cadence is what sharpens the suggestions, not a better prompt.
Phrase it as what changes, not as a promise:
Publish these and I'll have something to read next time. Once a few are out, I can look at what actually got replies and saves for this account and bring ideas shaped like the ones that worked, instead of ideas shaped like the topic.
| Published posts on the account | What this skill can offer |
|---|---|
| None | Topic-led ideas only — trends and the user's raw material |
| A few | Which formats the account has actually used, and gaps |
| Enough for a pattern | Ideas shaped like that account's own top performers |
Say it once per session, at the end, and never as a reason to publish now. Everything is still a draft, scheduling still counts as publishing, and the user's approval is still required. A nudge toward cadence must never read as pressure to approve the drafts on screen.
The reading itself is not this skill's job — postey-analytics owns
analytics.top_posts and post.analytics. If it is not installed, say the loop needs it rather
than implying this pack will do it.