Platform specs and validation
Skill social-media-skills/skills/skills/platform-specs-and-validation
106 social media skills for AI agents - strategy, writing, video, design, platform growth, publishing, and analytics. Works with Claude, Cursor, OpenClaw, Hermes & 40+ agents.
npx -y skills add social-media-skills/skills --skill platform-specs-and-validationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 23 days oldThe repository was created 23 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Platform specs and publish validation — per-platform requirements and the validate-before-publish step for multi-platform fan-out. Use when fanning a post out to multiple platforms, when a post is getting rejected, when mapping per-platform postType/required fields, or to run the validate-before-publish step. Encodes the per-platform postType enums, required-field matrices, and media rules from the OpenAPI spec, and runs POST /posts/validate so a post lands correctly on every target. Uses the CHECK framework. The keystone the per-platform format skills point to. WoopSocial validates + publishes atomically and has no update endpoint; the agent maps fields and fixes spec violations, but human-judgment calls (disclosure truthfulness, privacy intent) stay with the person; metrics never fabricated. Distinct from the per-platform format skills (one platform's content) and scheduling-and-queue (timing).
SKILL.md
6.4 KB, ~1.3k tokens by cl100k_base, as published. Nobody here has run it
platform-specs-and-validation
The cross-cutting publish-readiness skill — make a post correct on every target (required fields, media rules, limits) and run the validate-before-publish loop. It's the keystone the format skills point to. The agent maps fields + fixes spec violations + validates; WoopSocial validates and publishes; content craft and human-judgment calls stay elsewhere.
The POV: publish-ready, atomically, the first time
WoopSocial validates and publishes atomically — POST /posts checks every target together, and if
any one fails, nothing is created. There's no update endpoint, so a post that publishes wrong can't
be patched (edit = delete + recreate). That makes getting every target's fields right before
publish the whole game: the right postType, the required per-platform fields, the media rules, and
account-specific values (Pinterest board, TikTok privacy) resolved from platform-inputs — then
POST /posts/validate to catch what's left. This skill is where the format skills come to make sure it
will actually go through.
The framework: CHECK
(Depth: references/the-check-framework.md.)
- C — Collect the targets + project: list the social accounts; confirm they're in the same project; note each platform's field set.
- H — Hydrate per-platform required fields: fill each target's required fields (FB/IG
postType; TikTok's eight; YouTubetitle+privacy; PinterestpinterestBoardId); resolve account values fromGET /social-accounts/{id}/platform-inputs. - E — Enforce media + content rules: FB no image+video mix; IG Stories need media; Pinterest single
media; TikTok PHOTO/VIDEO field applicability; use
contentOverrideor separate posts where fan-out conflicts; respect text limits. - C — Confirm truthful declarations:
isBrandedContent/isYourBrand/madeForKidsare truthful declarations, not validation knobs — never flip them to clear an error; these stay with the person. - K — Knock it through validation:
POST /posts/validate→ fix all errors, review warnings → create (atomic). No update → delete + recreate (only whileNOT_STARTED).
The spec (ground truth — verify-quarterly)
Per-platform postType enums + required-field matrices + media rules + the platform-inputs discovery
endpoint + lifecycle facts (atomic create, no update, delivery status, schedule types):
references/woopsocial-publish-spec.md. The validate-then-create loop, the full ValidationField code
table + CreatePostErrorCode list + worked examples: references/validation-reference-and-recipes.md.
(The live OpenAPI spec / tools/integrations/woopsocial.md are the source of truth; enums and limits move.)
Honest scope (never violate)
- This skill ensures publish-READINESS (fields/media/spec correct + validated) — it doesn't write or improve the creative (the format/content skills do) and can't guarantee performance.
- WoopSocial validates + publishes atomically; one failing target fails the whole post. No update →
delete + recreate (only while
NOT_STARTED). - Disclosure/compliance declarations stay truthful (
isBrandedContent/isYourBrand/madeForKids) — never flipped to pass validation; privacy intent + disclosure truthfulness stay with the person. - Resolve account values (board, privacy) from
platform-inputs— don't guess. Never fabricate metrics — native analytics only. (Scope, distinctions + connections:references/scope-and-connections.md.)
Distinct from its siblings (route correctly)
platform-specs-and-validation (this) = the cross-platform field/media/validation layer that makes a post
publish-ready on every target · the per-platform format skills (tiktok-photo-mode,
tiktok-video-publishing, threads-post, youtube-publishing-and-metadata, carousel-writer, …) =
craft one platform's content and point here for validation · scheduling-and-queue = timing/cadence,
not correctness · tools/integrations/woopsocial.md = the shared integration guide this applies post-level.
Where this connects
Receives from: every format skill (crafted content + chosen postType) and content-calendar (the
plan). Resolves with: GET /social-accounts/{id}/platform-inputs (boards, privacy options) + the
integration guide. Hands to: scheduling-and-queue → WoopSocial (POST /posts after isValid),
then webhooks/delivery status for confirmation. Supports: cross-platform-repurposing (per-target
contentOverride), analytics-and-reporting (native delivery readout), experimentation (valid
variants). Content craft + truthful declarations stay with the format skills + the person.
Definition of done
Every target collected and confirmed in one project; each target's required fields hydrated (postType /
privacy / board / disclosure / title as applicable) with account-specific values resolved from
platform-inputs; media + content rules enforced (no FB mix, IG-Story media, single Pinterest media, TikTok
field applicability) with contentOverride or separate posts where fan-out conflicts; disclosure/madeForKids
declarations left truthful (never flipped to pass); POST /posts/validate run with all errors resolved and
warnings reviewed before an atomic create; edits handled via delete+recreate, nothing fabricated and no
promised performance; correctly distinguished from the per-platform format skills and scheduling-and-queue.
What ships with it: 5 files
17.8 KB alongside SKILL.md
evals/
- evals.json6.4 KB