메뉴
Convert approved articles into platform-optimized social posts. Multi-platform (Facebook, X/Twitter, LinkedIn, Bluesky, Threads). Legally defensible marketing copy—libel exposure is identical to article text. A/B test hooks. Metadata verified.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Transform raw OSINT into queryable databases. Ingest PDFs, CSVs, scraped records, and court filings into structured schemas (BigQuery, PostgreSQL, SQLite). Reproducible queries, not ad-hoc summaries. Chain of custody for evidence.
Use when exposing abuses of public trust, investigating political/corporate corruption, mapping financial data to policy outcomes, or pursuing suppressed public-interest truths where evidentiary rigorousness is the priority. Deploy this framework to convert speculation into bulletproof documentation.
Newsroom workflow air-traffic controller. Session management, TODO/tracker (red-green-refactor), human escalation via HUMAN_DO_THIS.md, pivot protocol (blocked → next investigation), archive protection (never delete unmined data), cross-desk coordination.
Adversarial pre-publication libel filter. Assume maximum liability. Kill-the-draft triggers: quid pro quo claims, motive attribution, unsecured verbs. Verification mapping (every financial assertion → primary source). Syntax conversion (definitive intention → observable contradiction). Publisher greenlight criteria.
Line-edit for readability, accessibility, and SEO without touching facts. This is your technical publishing QA before editorial review—catch typographical errors, rhythm problems, ad placement conflicts, and metadata gaps that confuse search engines.
Safe single-article deployment to production. File-by-file only (never wildcards/recursion). Pre-deployment checklist, exact file targeting, post-deployment verification, rollback procedure.
| name | social-distributor |
| description | Convert approved articles into platform-optimized social posts. Multi-platform (Facebook, X/Twitter, LinkedIn, Bluesky, Threads). Legally defensible marketing copy—libel exposure is identical to article text. A/B test hooks. Metadata verified. |
Social Distributor transforms a published, fact-checked article into platform-optimized social media posts that drive traffic while maintaining legal defensibility.
Critical principle: Marketing copy is legally liable to the same libel/defamation standards as the article itself. A misleading social hook can expose the newsroom to legal risk even if the article is accurate.
This skill covers:
Audience Profile:
Post Format:
| Element | Specification | Example |
|---|---|---|
| Hook (first 2 lines) | 2–3 sentences; emotional appeal; no link in text | "We spent 6 months investigating a hidden problem affecting families in your area. Here's what we found." |
| Body (context) | 3–5 sentences; summarize article, tease reveal; no spoilers | "Our team obtained internal documents showing..." |
| Image | 1200×628px (16:9 ratio); avoid text overlay; attractive/surprising | High-quality photo from article or custom graphic |
| Link | Shortened URL at end (bit.ly, news.co/...) | "Read full investigation: [link]" |
| CTA (call-to-action) | "Read more," "Learn what we found," "See the full investigation" (not "Click here") | "Read the full investigation" |
| Hashtags | 2–3 relevant (not #fb or platform meta-tags) | #Investigation #LocalNews #YourCity |
Character Limit: None (Facebook de-prioritizes longer posts but doesn't truncate)
Example Post:
We spent 6 months investigating a hidden problem in your community.
We found: [one-sentence revelation].
Our reporters obtained internal documents showing [tease]. Here's what you need to know:
• [Key finding 1]
• [Key finding 2]
• [Call to action: Read full investigation]
Read the full investigation: [link]
#Investigation #LocalNews
Audience Profile:
Post Format:
| Element | Specification | Example |
|---|---|---|
| Tweet 1 (hook) | 280 chars max; one headline-like fact; no link | "Breaking: Municipal contracts worth $5M went to single vendor with no competitive bidding." |
| Tweet 2–3 (thread) | Build narrative; each tweet <280 chars; link at end | "We obtained city council records showing [detail]. [Link to article]" |
| Image | Optional (reduces text character count but increases engagement) | Social card/screenshot |
| Link format | Shortened (t.co auto-shortens; use bit.ly for tracking) | Standard news site link |
Character Limit: 280 characters per tweet (including spaces, punctuation, links)
Threading Strategy:
Tweet 1 (hook):
Breaking: Municipal contracts worth $5M went to single vendor with no competitive bidding.
(137 chars)
Tweet 2 (data):
We obtained city council records revealing the vendor was awarded 47 of 50 contracts since 2020.
No other bidders were solicited.
(136 chars)
Tweet 3 (CTA):
Full investigation: [link]
Retweet Bait Tips:
Audience Profile:
Post Format:
| Element | Specification | Example |
|---|---|---|
| Opening hook | Professional tone; establish credibility/authority | "Our 6-month investigation into municipal procurement reveals systemic bias in contract awards." |
| Context/problem statement | 2–3 sentences; why this matters to professionals | "In government, competitive bidding ensures fair resource allocation. When skipped, taxpayers lose." |
| Data/insight | 3–5 bullet points; concrete findings | "• 94% of $5M in contracts awarded to one vendor • Zero competitive solicitations • Pattern consistent 2020–2024" |
| Takeaway | What should change; policy angle | "Governments must require competitive bidding documentation for all contracts >$50K." |
| Link | Full article link | "Read full analysis: [link]" |
| Image | Chart, graph, or data visualization (1200×628px) | Bar chart showing vendor concentration |
Character Limit: 3,000 characters (includes spaces, line breaks)
Example Post:
Our 6-month investigation into municipal procurement reveals systemic bias in contract awards.
In government, competitive bidding ensures fair resource allocation. When skipped, taxpayers lose.
Key findings:
• 94% of $5M in contracts awarded to single vendor
• Zero competitive solicitations since 2020
• Pattern identified across 4 fiscal years
• Internal documents show decision-makers did not follow procurement policy
Governments must require competitive bidding documentation for all contracts >$50K. Transparency protects taxpayers.
Read the full analysis: [link]
Audience Profile:
Post Format:
Similar to X/Twitter but:
Example Post:
We investigated municipal procurement over 6 months. Found: one vendor awarded 47 of 50 contracts worth $5M.
City council records show zero competitive bidding solicited. Either procurement policy wasn't enforced, or it was deliberately circumvented.
Full investigation: [link]
Audience Profile:
Post Format:
Treat like X/Twitter with Threads-specific features:
Example Thread:
Post 1: "We spent 6 months investigating municipal contracts. What we found should alarm every taxpayer."
Post 2: "One vendor was awarded $5M across 47 contracts since 2020. Competitive bids? Zero."
Post 3: "City council records obtained by our reporters show procurement policy was either unenforced or deliberately bypassed."
Post 4: "The real question: was this negligence or corruption? Read our investigation: [link]"
| Hook Type | Pattern | Example | ⚠️ Legal Risk | ✅ Safe Version |
|---|---|---|---|---|
| Surprising Stat | "[Number]% of [X] shows [unexpected pattern]" | "94% of city contracts awarded to one vendor—zero competitive bids." | Low (if stat verified) | Same; fact from article |
| Question | "Did [decision-maker] know about [consequence]?" | "Did city council know competitive bidding was being skipped?" | MEDIUM (implies knowledge; avoids "deliberately" but hints at motive) | "City council records show competitive bidding was skipped. Why?" (observable fact, not motive) |
| Local Paradox | "[City/org] claims [principle]. Reality: [contradiction]." | "City claims fair procurement. Reality: one vendor got 94% of contracts." | Low (compares claimed vs. actual) | Same |
| Consequence Angle | "[X] means [Y impact] for [audience]." | "No competitive bidding means less oversight. Taxpayers lose." | Low (causal reasoning) | Same; state as analysis, not certainty |
| Human Impact | "[Group] affected by [finding]" | "5,000 families impacted by hidden water contamination—here's why." | Low (if specific, verifiable) | Confirm number before posting |
| Investigative Process | "We obtained [source]; here's what it shows:" | "We obtained 3 years of council records. Procurement is broken." | Low (if source real) | Same; link to source |
| Phrase | Why It's Risky | Safe Alternative |
|---|---|---|
| "City officials deliberately hid [fact]" | Attributes motive without stating evidence of intent | "City officials did not disclose [fact]" (observable action) |
| "Corruption uncovered" | Legal term; implies criminal intent (defamatory) | "Irregularities found in [specific process]" |
| "Mayor took bribes" | Explicit accusation; defamatory if unproven | "Mayor accepted $X from vendor; no disclosure filed" (observable fact) |
| "Taxpayers defrauded" | Assumes illegal intent; implies crime | "Taxpayers paid more due to [specific failure]" |
| "Cover-up" | Implies intent to hide; requires proof of motive | "Documents were not publicly disclosed" (observable) |
| "Mayor's cronies" | Character attack; not fact-based | "Contracts awarded to former mayor's business partner" (relationship is fact) |
Test 2–3 hooks for the same article; measure click-through rate (CTR):
article: "Municipal Procurement Investigation"
publish_date: "2025-04-01"
platform: "Facebook"
hook_a:
text: "We spent 6 months investigating city contracts. What we found is shocking."
post_date: "2025-04-01 09:00 AM"
reach: 5420
clicks: 127
ctr: "2.3%"
hook_b:
text: "94% of city contracts went to one vendor. Zero competitive bids. Here's how it happened."
post_date: "2025-04-01 11:00 AM"
reach: 6100
clicks: 289
ctr: "4.7%"
hook_c:
text: "Your tax dollars: $5M to one vendor, no questions asked. Our investigation explains why."
post_date: "2025-04-01 02:00 PM"
reach: 5900
clicks: 201
ctr: "3.4%"
winner: "hook_b"
insights: "Specific data (94%, one vendor) outperforms vague language (shocking, questions asked)"
Repeat: Repost winning hook to other platforms with format adjustments.
When article is shared on Facebook, X, LinkedIn, the platform reads OpenGraph (OG) metadata:
| OG Tag | Purpose | Specification |
|---|---|---|
og:title | Headline for social card | 50–65 chars; same or variant of article headline |
og:description | Summary for social card | 155–160 chars; compelling hook, not just first paragraph |
og:image | Thumbnail image | 1200×628px (16:9); JPG/PNG; <200KB; unique per article |
og:type | Content type | "article" for news |
og:url | Canonical URL | Full article URL; no tracking parameters in OG tag |
<!-- Example OG tags in article HTML -->
<meta property="og:title" content="Municipal Contracts: One Vendor Received 94% of $5M Awards">
<meta property="og:description" content="Our investigation reveals systemic procurement bias. City council records show zero competitive bidding over 4 years.">
<meta property="og:image" content="https://newsroom.org/images/article-2025-04-01-contracts.jpg">
<meta property="og:url" content="https://newsroom.org/articles/2025-04-01-procurement-investigation.html">
<meta property="og:type" content="article">
Test: Use Facebook Sharing Debugger (facebook.com/developers/tools/debug) to preview card.
| Platform | Limit | Strategy | Example |
|---|---|---|---|
| X/Twitter | 280 chars | Brutal concision; use numbers; thread for context | "94% of $5M contracts to one vendor. Zero competitive bids." (67 chars) |
| None | 2–3 sentences + link + image (length doesn't matter, but ~200 chars for preview) | "We spent 6 months investigating municipal contracts. One vendor got 94%. We obtained the proof." | |
| 3,000 chars | Use full allowance for context, data, analysis | Full paragraph + bullets + insight | |
| Bluesky | ~300 chars | Slightly more than X; allow nuance | "Our investigation: municipal contracts worth $5M awarded to single vendor. City claims fair procurement. Records show zero competitive bidding." |
| Threads | 500 chars per post | Use threads for narrative; each post ~500 chars | 5–10 connected posts, each building story |
Optimization Rules:
A social post is defamatory if it:
Marketing copy can be defamatory. Your headline can be accurate, but the social hook can be inaccurate or misleading.
| Risky Phrase | Safe Phrase | Why It's Safer |
|---|---|---|
| "Mayor took bribes" | "Mayor received $X from vendor; undisclosed in filings" | States observable facts; avoids criminal accusation |
| "Corruption at city hall" | "City procurement shows irregular spending patterns" | Describes finding; avoids legal-adjacent language |
| "Official lied to public" | "Official's statement contradicts released documents" | Compares claim to evidence; doesn't attribute intent |
| "Embezzlement investigation" | "City treasurer misappropriated $X; investigation ongoing" | Describes action; avoids legal conclusion |
| "Conspiracy uncovered" | "Three officials coordinated decisions affecting vendor selection" | Describes observable behavior; avoids conspiracy accusation |
Repost winning hook to all platforms (adjusted for character limits):
Tag relevant accounts (journalists, organizations, decision-makers mentioned in article—sparingly, not spam)
Engage with replies (respond to comments; answer reader questions; cite article for fact-checks)
| Mistake | Why It's Bad | Fix |
|---|---|---|
| New claim in hook not in article | Defamatory if false; readers click expecting to see evidence they won't find | Every social claim must appear verbatim (or close paraphrase) in published article |
| Clickbait headline ("This One Trick Will Shock You") | Low-quality audiences click; high bounce rate; undermines credibility | Use specific fact from article as hook; avoid hype |
| Motive attribution ("Mayor deliberately bypassed procurement to help friend") | Defamatory if motive unproven; states criminal intent | Rephrase: "Mayor bypassed procurement; vendor is campaign donor [cite]" (facts, not motive) |
| Hedge language ("Possibly the largest contract award...") | Weakens authority; readers unsure if true; poor CTR | If verified: "Is the largest contract..."; if uncertain, don't post yet |
| No specific detail ("City hall in crisis") | Too vague; no reason to click; looks like sensationalism | Add fact: "97% of contracts to one vendor; no competitive bids" |
| Call-out/tag every person mentioned | Spam; looks aggressive; may trigger harassment reports | Tag only directly quoted/interviewed sources; max 2–3 tags per post |
| Mistake | Why It's Bad | Fix |
|---|---|---|
| og:image is generic placeholder | Social card looks low-quality; low CTR; appears lazy | Create unique social card per article; use article photo or custom graphic |
| og:image is >200KB | Slow to load; may not display; wasted bandwidth | Compress to <200KB (use TinyPNG); maintain 1200×628px ratio |
| og:title is click-bait | Misleads readers; high bounce rate; damages publication credibility | og:title should match or closely reflect article headline |
| og:description is first paragraph copy-pasted | Boring; no hook; poor preview | Write compelling og:description (155–160 chars) that entices click |
| og:url includes tracking parameters | OG tag should be canonical; may break social caching | Use clean article URL without utm_source, etc. (add tracking to actual posted link separately) |
| Mistake | Why It's Bad | Fix |
|---|---|---|
| Only post one hook; declare "winner" based on gut feeling | No data; subjective; can't replicate success | Always test 2–3 hooks; measure CTR; declare winner after 2 hours |
| Post all hooks at same time | Can't isolate which hook performs; confounding variables (time of day) | Stagger hooks by 1–2 hours; control for timing |
| Test on different platforms ("Facebook wins, so Twitter loses") | Audiences different; can't compare X hook to Facebook hook | Test different HOOKS on SAME platform; compare apples to apples |
| Stop after first test; never repeat | One win ≠ pattern; may have been timing luck | Run 2–3 A/B tests over different weeks; identify pattern |
| Mistake | Why It's Bad | Fix |
|---|---|---|
| Ignore comments/replies | Readers feel unheard; damages relationship; looks dismissive | Reply to first 10–20 substantive comments; thank shares |
| Argue with critics in replies | Escalates conflict; looks unprofessional; algorithm downranks | Reply once (factually); don't continue debate |
| Promote without fact-checking replies | Comments may contain misinformation; newsroom doesn't correct | Pin your comment with fact-check if significant inaccuracy in top reply |
| Repost at same time every day | Algorithm learns pattern; deprioritizes | Vary posting time; post when audience is most active (use platform analytics) |
Before posting any social hook:
If a social post receives cease-and-desist or legal threat:
| Platform | Best for | Max Hooks | Posting Frequency | Image Requirement |
|---|---|---|---|---|
| Emotional, narrative | 3–5 | 1–2x daily | Required (1200×628px) | |
| X/Twitter | Breaking, timely, threads | 5+ (thread) | 2–4x daily | Optional (helps) |
| Professional, analysis, B2B | 2–3 | 3–5x weekly | Recommended (charts) | |
| Bluesky | Engaged, early-adopter | 2–3 | 1x daily | Optional |
| Threads | Narrative, long-form | 1 (as thread) | 2–3x weekly | Optional (builds engagement) |