Shaduf.Research preview
Marketing Skills Catalog/One approved article to two social drafts: source boundary

Research report · 27 October 2026

One approved article → two social drafts

Core Haines Social Content describes turning a blog insight into LinkedIn copy and takeaways into an X thread. That supports a narrow drafting hypothesis, not a fifth catalogue card or a posting recommendation.

The bounded job

Given one already-approved, non-sensitive article, produce one LinkedIn text draft and one short X thread draft as local, unscheduled text. The brief would need the article's approved facts, canonical URL, attribution and rights, intended reader, voice and exact CTA. None was supplied for this review. Each material claim should point to an article passage or be marked unresolved. A named reviewer checks factual fidelity, quotations and rights, voice, URL and CTA, then separately decides whether to publish.

The claim pointers and local release gate are proposed review controls, not requirements found in the skill source. “Short” is a proposed output constraint: the source templates show roughly seven- or eight-post X examples, not a short-thread rule.

Source instruction versus this job

Inspected sourceSource describesBoundary for two local drafts
Social Content selected fileRead product-marketing context, ask about goal, audience and voice, extract article ideas, and adapt a blog insight for LinkedIn and takeaways for X.Draft formats are source-supported. Local context reads need separate authorization. Choosing an angle does not create new evidence.
Publishing and engagement sectionsCheck connected tools, create remote drafts, request scheduling approval, schedule, respond and later read statistics.Remote drafts, accounts, scheduling, replies and monitoring are outside this local job. Source-described approval does not establish host controls.
Platform guide and limits referenceFormat, cadence, timing, link-placement, reach and character guidance.Use as dated drafting context only. Neither current platform policy nor performance was verified; a publishing reviewer must check applicable rules.
Post templates and X algorithm noteGeneric structures, result and CTA placeholders, and a May 2026 interpretation whose weights are withheld and numbers may age.Do not invent outcomes or fill a CTA placeholder without approval. Some template engagement questions conflict with the main file's anti-bait advice. No reach prediction follows.

What would change this decision?

A stable selected revision, complete same-skill payload and chosen-host permission surface would be needed before a controlled comparison. With an authorized article, compare the template and skill route using the same brief and fixed rubric: traceable claims, quotation/rights and CTA errors, unauthorized reads or actions, reviewer effort and useful draft quality. If the template matches or beats the skill without material omissions, the added-value hypothesis fails.

The broader source also contains calendar, analytics, video and optimization claims. They were not adopted or outcome-tested for this text-draft decision.

Provenance and limits

The mutable main selected file displayed name: social and version 2.4.0 in a web-broker rendering on 27 October. It and four linked same-skill references were read. “Crawled: today” and “Crawled: 2 days ago” were broker cache labels, not proof of fresh origin content. A cached path-history page did not align with the displayed file; API access and container Git resolution failed. No selected-file SHA/date, immutable intended revision, full package inventory, installed payload or effective host behavior was established.

No article, account, analytics, synthetic case, skill invocation, scheduling, posting or marketing outcome was used or observed. Source links describe instructions and claims, not a verified result.

Research library · Four active candidates · Original selected file

Search published pools, pages, reports, and evidence.