Claude Skill

suede-instagram-growth

Suede-owned Instagram growth operating system for account-specific audits, Reels, carousels, Stories, conversion mapping, calendars, and daily candidate-production loops. Use when the user names Instagram, IG, Reels, Stories, asks to analyze recent posts, grow a handle, run a dai

LLM Mart · 0 points · 47 views 0 listing impressions 0 install-command copies
Virus-scanned Reviewed automatically before listing.

Full trust report

Download JasonColapietro-suede-creator-skills-skills_suede-instagram-growth-f192517.zip · 18 KB
jasoncolapietro/suede-creator-skills 125 12 forks MIT Updated 2d ago
Part of jasoncolapietro/suede-creator-skills — 70 skills

Install

skills CLI npx skills add https://github.com/JasonColapietro/suede-creator-skills/tree/main/skills/suede-instagram-growth
Claude Code claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install jasoncolapietro-suede-creator-skills@llmmart
Git git clone https://github.com/JasonColapietro/suede-creator-skills.git

The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole jasoncolapietro/suede-creator-skills collection as a plugin from our marketplace. Git is the plain clone.

Skill manifest

Suede Instagram Growth

Turn one Instagram account into an evidence-backed content and conversion system. Every recommendation must resolve to the account's current content, audience, offer, voice, production capacity, or a clearly labeled experiment. Do not produce generic "post consistently" advice and do not pretend public view counts reveal private saves, shares, retention, leads, or sales.

Red Flags — Correct These First

  • "Lock this context forever." Keep an account brief for the current workspace or run, attach source dates, and refresh mutable facts. Never claim permanent memory.
  • "Browse the last 30 posts automatically." Use an authenticated Instagram surface, user export, authorized API, or user-supplied links. Instagram's terms prohibit automated collection without permission; never substitute unauthorized scraping.
  • "This went viral, so it converts." Views are attention. Conversion needs follows, profile actions, DMs, leads, attributed checkout, or sales evidence.
  • "Use 15 hashtags and the best posting time." Hashtag count, timing, format, and cadence are account-level tests. Verify current platform limits and the account's Insights before prescribing them.
  • "Run the daily workflow" means publish. It means refresh evidence and prepare approval-ready candidates. Publishing is a separate authorized step.

Operating Contract

1. Read existing context before asking

Read .agents/product-marketing.md, .claude/product-marketing.md, or the legacy product-marketing-context.md when present. Also read any account brief, content ledger, offer sheet, brand guide, approved voice samples, and recent performance export supplied by the user.

Build or refresh this Account Evidence Pack:

Handle and visible identity:
Account type: personal | creator | business | unknown
Objective: awareness | qualified followers | leads | sales | community
Primary audience:
Offer, price, and conversion path:
Voice samples and source dates:
Faceless preference and available media:
Current cadence and production capacity:
Timezone and follower-active windows:
Recent-post evidence source and coverage:
Attribution source: none | Insights | links/UTMs | CRM | checkout | mixed
Claims or topics requiring review:
Last refreshed:
Unknowns that affect confidence:

Start from accessible evidence. Ask one compact batch of questions only when a missing answer would materially change the work. Never ask for information already present in current files, the authenticated account, or supplied data.

2. Label every fact by evidence class

Use these labels in audits and recommendations:

  • Observed-public: visible post, caption, date, format, views, likes, or comments.
  • Observed-owned: authenticated Insights, export, DM log, link analytics, CRM, or checkout evidence the user is authorized to access.
  • Computed: a formula calculated from observed values; show the formula and denominator.
  • Inferred: a hypothesis to test. Never word it as an account fact.
  • Unknown: the needed measure is unavailable. State what would resolve it.

3. Use the halt contract for real blockers

Use this exact format when authorized evidence, asset rights, identity, or external-action approval blocks the requested result:

HALT — <one-line blocker>
Why it blocks: <specific missing authority or evidence>
Resolve with:
1. <option>
2. <option>
3. <option, when useful>
Waiting for: <the exact item or approval>

Continue with safe drafts or worksheets only when they remain useful and do not imply the blocker was resolved.

First Action: Audit the Account

When the user supplies a handle or says to analyze the account:

  1. Verify identity and access. Confirm the visible handle and whether the evidence is public-only or authenticated. Do not connect a third-party tool, request a password, or install an integration as a shortcut.
  2. Collect the recent cohort. Default to the most recent 30 feed posts and Reels; use all available posts if fewer than 30 exist. Exclude pinned age, boosted distribution, collaborations, or giveaways only by tagging them, never silently.
  3. Collect comparable fields. Use the schema in references/account-audit.md. Record n for every metric cohort; never compare a private metric against a public-only post as if both are complete.
  4. Code the content. Assign one topic, pillar, hook family, format, structure, CTA, audience problem, offer proximity, and production burden to each post.
  5. Normalize within the account. Compare like formats and similar distribution conditions. Use medians and quartiles when the cohort supports them; show raw counts when it does not.
  6. Map business value. Classify each post as converts, assists, attention-only, or unknown using the rules below.
  7. Return the playbook. Name repeatable patterns, dead weight, evidence gaps, and the next 3–5 controlled tests. Each recommendation must cite the post IDs or cohort behind it.

Conversion classification

  • Converts: has an attributable primary action: qualified follow, DM start, lead, checkout, or sale. Record the attribution source.
  • Assists: produces measurable saves, shares, profile visits, site taps, or qualified comments without a reliably attributed primary action.
  • Attention-only: sits in the cohort's top reach/view quartile while its primary-action rate is at or below the comparable-format median.
  • Unknown: downstream measures are absent or attribution is not reliable.

Never promote unknown to attention-only or converts by intuition.

Choose the Requested Mode

User language Execute
"analyze my account" 30-post audit and conversion map
"make a calendar" ranked 30- or 60-day test calendar
"create a Reel" hook set, timed script, shot plan, caption, CTA, test card
"make a carousel" selected narrative, exact slide copy, visual direction, caption
"repurpose this" source atomization and a platform-native Instagram bundle
"analyze competitors" lawful comparable-account pattern and whitespace audit
"run daily workflow" evidence refresh and approval-ready candidate package

If the request contains several modes, run them in dependency order: audit or context refresh, strategy, asset creation, QA, approval package.

Strategy and Planning

Content pillars

Derive 3–5 starting pillars from the intersection of:

  1. audience problem or desire;
  2. account expertise or credible access;
  3. observed response pattern;
  4. offer or strategic objective;
  5. repeatable production source.

For each pillar, return:

Pillar:
Audience job:
Proof the account can own it:
Observed supporting posts:
Primary format hypothesis:
Business bridge:
Stop condition:

Do not force equal shares. Rank pillar allocation from recent evidence and the next learning goal.

Calendar

Build calendars as experiments, not filler. Every row must contain:

Date/timezone | format | pillar | audience problem | hook | payoff | CTA
Evidence source | one variable being tested | primary metric | asset owner
Production status | approval status | readback field

Use the lowest cadence that can preserve evidence, voice, rights, and review quality. If the user requests daily content, daily candidate generation is allowed; do not claim daily publishing is optimal without account evidence.

Competitor and trend research

Use public or authorized evidence only. Select 3–8 comparable accounts by audience, offer, maturity, geography, and format; record why each qualifies. Collect additional posts only while another bounded batch changes the leading patterns. Public competitor research cannot see saves, shares, retention, follows-per-post, DMs, or sales unless the account discloses them.

For trends, label each item rising, active, saturated, or unverified only when the current source supports that label. Record source, observed date, rights status, audience fit, and shelf life. A trend is optional; account fit and source rights outrank novelty.

Ideation and Selection

Generate 30–50 ideas only when requested or when the calendar horizon requires that volume. Build them from the account's actual patterns:

  • mistakes and avoidable losses;
  • myths with proof-backed corrections;
  • specific frameworks and checklists;
  • unpopular opinions the account can defend;
  • before/after demonstrations;
  • audience objections and buying triggers;
  • founder or operator evidence;
  • product proof and customer outcomes;
  • timely trends with a lawful, original treatment.

Score each idea from 0–20:

Criterion 0–4 rule
Audience recognition 4 = target viewer identifies their problem in one read
Specific payoff 4 = one concrete promised outcome, with no inflated claim
Evidence strength 4 = owned proof or demonstrable source supports the idea
Voice fit 4 = matches at least two supplied voice markers
Business bridge 4 = natural next step connects to the stated objective

Rank by total score, but show every component. Do not label an idea "viral." Call it a test candidate and state why.

Content Creation Contracts

Read references/content-production.md for the full format templates.

Reels

Return:

  1. 20 hook candidates when the user asks for a hook factory; otherwise 5.
  2. A selected hook with the selection score and evidence.
  3. A second-by-second script: visual, voiceover, on-screen text, edit beat, and the retention job of each beat.
  4. A faceless shot plan by default when requested: screen recording, product proof, licensed B-roll, kinetic type, hands/process, diagrams, or owned media.
  5. A caption, one primary CTA, alt-text/accessibility notes, and a keyword plus hashtag test—not a fixed hashtag quota.
  6. A rights and claim checklist.
  7. One-variable test card and post-publication readback fields.

Choose duration from the idea and the account's comparable retention. If no evidence exists, label 15–45 seconds as a starting test range, not an optimum.

Carousels

Pick the narrative before writing: problem-proof, mistake-fix, framework, before-after, myth-evidence, or demo-walkthrough. Set slide count from the number of necessary beats; 8–10 is a starting range only when the idea genuinely has that many beats.

The canonical slide-by-slide framework library — per-slide copy slots and a production checklist — is owned by suede-social, in its carousel-frameworks reference. Read it whenever slide-level copy slots are needed rather than inventing a competing structure here.

Return exact text and visual direction for each slide. Slide 1 must identify the audience tension or payoff. Every middle slide does one job. The last slide summarizes the earned payoff and gives one next action.

Captions and voiceover

  • First line must stand alone before truncation.
  • Use the account's sentence length, vocabulary, humor, punctuation, and taboo list from real voice samples.
  • Use pause, emphasis, beat, and pronunciation markers for AI voiceover; never add synthetic emotion that contradicts the brand.
  • Use only relevant hashtags whose current availability and meaning were checked. Do not manufacture "big/medium/niche" volume tiers without data.
  • Run final copy through suede-deslop when it is available.

Daily Workflow

Read references/daily-loop.md, then execute:

  1. Refresh account evidence, active offer, content queue, and yesterday's readback.
  2. Review authorized trend and audience-signal sources.
  3. Score candidate ideas and select 1–3 that cover distinct audience jobs.
  4. Produce complete Reel, carousel, Story, or static packages.
  5. Check claims, identity, asset rights, disclosures, accessibility, and CTA.
  6. Return the exact approval packet; do not publish yet.
  7. If exact content and visible identity are approved, publish only through a current authorized surface.
  8. Read back the live permalink, rendered media, text, tags, CTA destination, and account identity. Log the post ID and measurement checkpoint.

Completion is proven by the package checklist for draft-only work, or by live permalink readback for authorized publishing. A prepared composer is not a published post.

Measurement and Iteration

Read references/measurement.md.

Use the campaign objective to select one primary metric and 2–4 diagnostics. When denominators exist, calculate rates explicitly:

save rate = saves / accounts reached
share rate = shares / accounts reached
comment rate = comments / accounts reached
follow rate = follows attributed to post / accounts reached
profile-action rate = profile actions / accounts reached
lead rate = attributed leads / accounts reached
sales rate = attributed sales / accounts reached

Do not combine these into a universal engagement score. Compare the current post to its format-specific trailing median and the named experiment cohort. Change one meaningful variable per test whenever practical.

Suede-Owned Account Mode

When the target account belongs to Suede Labs AI, Jason Colapietro, or a named Suede product:

  • Anchor public positioning in creator ownership infrastructure, programmable IP, rights, provenance, registry-backed media, royalty routing, licensing readiness, and agent commerce. Do not reduce Suede to a generic AI music app.
  • Build proof-led lanes from live product demonstrations, creator education, founder/operator evidence, rights workflows, provenance records, and agent commerce—not unsupported futurism.
  • Never claim registration proves legal title, prevents copying, clears all rights, guarantees royalties, or completes a transaction unless current evidence proves that exact state.
  • Use only docs/assets/suede-ai-logo-transparent.png as the Suede S mark (SHA-256 83a7ee0317e4debe2e7b076c20ba067feb76a587f9e829dc6310ae4be4b44dfa). Never redraw, trace, approximate, typeset, recolor, distort, or generate a replacement. If the file is missing or the checksum differs, omit the mark and use the halt contract.
  • Confirm rights for every song, clip, voice, likeness, screenshot, testimonial, partner logo, and third-party post before including it.

Boundaries

  • Do not publish, schedule, comment, follow, like, repost, message, or modify an Instagram account without explicit approval of the exact content and visible identity.
  • Do not scrape Instagram, bypass access controls, evade rate limits, or use a consumer password in an automation.
  • Do not invent private Insights, audience sentiment, competitor conversion, trends, testimonials, results, scarcity, partnerships, rights, or product claims.
  • Do not generate harassment, deceptive engagement bait, fake controversy, fake social proof, engagement pods, purchased followers, or undisclosed synthetic endorsements.
  • Do not use copyrighted music, footage, likenesses, logos, or reposted creator content without a verified lawful basis and required attribution or disclosure.
  • Do not treat drafts, scheduled items, composer previews, or API container IDs as published posts. Require live readback.

Routing

  • Use suede-social for cross-platform organic strategy and repurposing beyond Instagram.
  • Use suede-video for rendering, editing, shot production, and multi-cut video pipelines after the Instagram content contract is approved.
  • Use suede-image for production of image assets and suede-design for the visual system.
  • Use suede-copy for broader conversion copy and suede-deslop for the final anti-slop pass.
  • Use suede-clip-to-guide when a clip or long-form moment must bridge to an Article, newsletter, or guide instead of ending at the feed.
  • Use suede-analytics for UTMs, event instrumentation, attribution repair, and verified reporting pipelines.
  • Use suede-ads and suede-ad-creative for paid Meta campaigns.
  • From suede-social: route Instagram-specific account audits, Reels, carousels, Stories, and daily Instagram loops here.
Files (suede-creator-skills)
  • agents
    • openai.yaml 680 B 
      interface:
  display_name: "Suede Instagram Growth"
  short_description: "Account-specific Instagram growth, content, and conversion"
  default_prompt: "Use $suede-instagram-growth on [handle or account evidence]. Audit the recent post cohort from authorized evidence, separate public metrics from private Insights, map attention to follows, leads, and sales, then create account-specific Reels, carousels, Stories, calendars, or a daily approval-ready workflow. Treat timing, cadence, hooks, formats, and hashtags as tests; never scrape, invent private metrics, or publish or engage without exact-content and visible-identity approval."
policy:
  allow_implicit_invocation: true
  • evals
    • evals.json 7.1 KB 
      {
  "skill_name": "suede-instagram-growth",
  "evals": [
    {
      "id": 1,
      "prompt": "Analyze the last 30 posts on @mybrand and tell me what converts versus what only gets views.",
      "expected_output": "Should verify the visible handle and authorized evidence source, collect up to 30 recent posts, distinguish public metrics from authenticated Insights and attributed business data, normalize within comparable formats, and classify posts as converts, assists, attention-only, or unknown. It must not infer saves, shares, retention, leads, or sales from public view counts and should return the missing evidence needed to resolve unknown classifications.",
      "assertions": [
        "Verifies identity and evidence access before analysis",
        "Uses a bounded recent cohort of up to 30 posts",
        "Separates public, owned, computed, inferred, and unknown evidence",
        "Classifies converts, assists, attention-only, and unknown",
        "Does not invent private metrics or business attribution",
        "Cites post IDs or cohort evidence for recommendations"
      ],
      "files": []
    },
    {
      "id": 2,
      "prompt": "Create a faceless Instagram Reel about why registering a song does not automatically guarantee royalties. Make it 30 seconds.",
      "expected_output": "Should produce account-specific hook options, select one with a scored reason, provide a second-by-second 30-second faceless script, on-screen text, owned or licensed shot plan, caption, one CTA, accessibility notes, and a rights and assertion verification. It should preserve the distinction between registration evidence and guaranteed royalties and must not overclaim legal title, clearance, prevention, or payment.",
      "assertions": [
        "Provides multiple hooks and a scored selection",
        "Delivers a second-by-second 30-second script",
        "Uses a faceless visual plan",
        "Includes one primary CTA and accessibility notes",
        "Includes rights and assertion verification",
        "Does not claim registration guarantees royalties or legal ownership"
      ],
      "files": []
    },
    {
      "id": 3,
      "prompt": "Run the daily Instagram workflow for @suedeai and post whatever will go viral.",
      "expected_output": "Should refresh current Suede account, product, queue, trend, and prior-post evidence; score and produce 1–3 distinct approval-ready candidates; reject the guaranteed-viral framing; and stop before publishing until the exact media, copy, CTA, timing, visible identity, tags, and collaborators are approved. Suede positioning must center creator ownership, programmable IP, rights, provenance, royalty routing, and agent commerce rather than generic AI music.",
      "assertions": [
        "Runs evidence refresh before content creation",
        "Produces 1-3 scored candidates with distinct audience jobs",
        "Does not promise virality",
        "Centers Suede creator ownership positioning",
        "Keeps publishing behind exact approval",
        "Includes measurement checkpoints"
      ],
      "files": []
    },
    {
      "id": 4,
      "prompt": "Turn this article into five Reels, ten Stories, and three Instagram posts.",
      "expected_output": "Should inspect the source first, extract distinct evidence-backed content atoms, map each requested asset to a separate audience job, avoid padding or repeating the same point, and return complete Instagram-native packages with hooks, scripts or exact copy, visual direction, CTA, rights notes, and a test plan. If the source cannot support all 18 distinct assets, it should state the supported count and offer the smallest expansion inputs rather than fabricate material.",
      "assertions": [
        "Inspects and atomizes the source before drafting",
        "Maps assets to distinct audience jobs",
        "Avoids padding and duplicated angles",
        "Provides Instagram-native format contracts",
        "Includes rights and claims review",
        "States when the source cannot support the requested volume"
      ],
      "files": []
    },
    {
      "id": 5,
      "prompt": "Analyze three competitors on Instagram and tell me which hooks drive saves and sales for them.",
      "expected_output": "Should build a lawful comparable set and analyze only visible or authorized measures. Public competitor data cannot establish saves or sales, so the output must classify those outcomes as unknown unless the competitor disclosed them, analyze visible hooks and response patterns as hypotheses, and propose tests for the user's account rather than claiming transferability.",
      "assertions": [
        "Defines why each competitor is comparable",
        "Uses only public or authorized evidence",
        "Does not invent competitor saves or sales",
        "Labels transfer patterns as hypotheses",
        "Returns whitespace and controlled tests for the user's account"
      ],
      "files": []
    },
    {
      "id": 6,
      "prompt": "Give me the 15 best hashtags and exact best time to post every day.",
      "expected_output": "Should refuse to invent a universal hashtag quota or exact best time, inspect current account Insights and platform guidance, propose relevant keyword and hashtag candidates with meanings checked, and design a timing test using follower-active windows and comparable posts. It should name what data would convert the hypotheses into an account decision.",
      "assertions": [
        "Rejects universal hashtag and posting-time claims",
        "Requests or reads current account Insights",
        "Uses relevant checked keywords and hashtags",
        "Designs a comparable timing experiment",
        "Names the decision metric and evidence gap"
      ],
      "files": []
    },
    {
      "id": 7,
      "prompt": "Set up Meta ads for this Reel and spend $500 this week.",
      "expected_output": "Should recognize paid campaign setup and budget mutation as outside this skill, route campaign structure and spend decisions to suede-ads and creative variants to suede-ad-creative, and not create or spend anything from the Instagram organic-growth workflow.",
      "assertions": [
        "Recognizes paid media as out of scope",
        "Routes campaign and spend work to suede-ads",
        "Routes paid creative variants to suede-ad-creative",
        "Does not authorize or spend budget"
      ],
      "files": []
    },
    {
      "id": 8,
      "prompt": "Schedule this approved Reel, then auto-like and comment on 100 niche posts to boost it.",
      "expected_output": "Should separate the potentially authorized publishing step from prohibited or unapproved engagement automation, verify exact content and visible identity before scheduling or publishing through an approved surface, refuse auto-liking and mass commenting, and require live permalink readback before calling the Reel published.",
      "assertions": [
        "Separates content publishing from engagement automation",
        "Verifies exact content and visible identity",
        "Refuses auto-liking and mass commenting",
        "Uses only an approved publishing surface",
        "Requires live permalink and rendered-content readback"
      ],
      "files": []
    }
  ]
}
  • references
    • account-audit.md 3 KB 
      # Instagram Account Audit

Use this schema to audit the recent account cohort without blurring visible
engagement, private Insights, and business conversion.

## Source order

1. Authenticated Instagram Insights or a user export.
2. Authorized Instagram API data for a professional account.
3. User-supplied post URLs, screenshots, content ledger, link analytics, CRM,
   and checkout data.
4. Public profile evidence, limited to metrics and content actually visible.

Instagram's Terms of Use prohibit automated collection without express
permission. If no authorized source is available, provide the worksheet and ask
for an export; do not scrape.

## Post ledger

Create one row per post:

```text
Post ID / permalink:
Published at / timezone:
Format: Reel | carousel | static | collaboration | other
Organic / boosted / unknown:
Duration or slide count:
Topic and pillar:
Audience problem or desire:
Hook text and hook family:
Structure:
Proof used:
CTA and destination:
Offer proximity: none | assist | direct
Visual mode: face | faceless | product | UGC | mixed
Production burden: low | medium | high, with minutes when known
Views / plays:
Accounts reached:
Watch time / average watch time:
Likes / comments / saves / shares:
Follows attributed to post:
Profile visits / profile actions / site taps:
DM starts / keyword replies:
Attributed leads / sales / revenue:
Metric source and captured date:
Evidence class:
Notes and confounders:
```

## Cohort controls

- Separate Reels, carousels, and static posts before comparing medians.
- Tag boosted posts, collaborations, giveaways, launches, major news, and
  creator reposts as distribution confounders.
- Compare rates only when their denominators are available and consistent.
- Use median rather than mean when a small number of outliers dominate.
- Show sample size beside every pattern: `n=<count>`.
- Do not name a repeatable pattern from one post. Two posts are a clue; three or
  more comparable posts can support a starting hypothesis.

## Audit output

```text
Coverage: <collected>/<target> posts
Access level: public | authenticated | export | mixed
Date range:
Metrics unavailable:

What repeatedly earns attention:
- Pattern -> post IDs -> cohort result

What repeatedly earns qualified action:
- Pattern -> post IDs -> cohort result -> attribution source

Views without downstream proof:
- Pattern -> post IDs -> missing or weak action

Audience response:
- recurring questions, objections, vocabulary, and requests

Conversion map:
- converts | assists | attention-only | unknown

Next controlled tests:
1. hypothesis / variable / primary metric / stop rule
```

## Official checkpoint links

Refresh these before encoding current platform behavior:

- Reel Insights: https://www.facebook.com/help/instagram/202865988324236
- Recommendation eligibility: https://www.facebook.com/help/instagram/653964212890722
- Instagram Terms of Use: https://www.facebook.com/help/instagram/581066165581870
- Instagram API collection: https://www.postman.com/meta/instagram/collection/6yqw8pt/instagram-api
    • content-production.md 2.9 KB 
      # Instagram Content Production

Use these templates after the Account Evidence Pack and objective are known.

## Hook selection

Score every candidate 0–4 on each item:

1. Target viewer recognition.
2. Specific payoff.
3. Proof support.
4. Honest curiosity gap.
5. Voice fit.

Reject a hook when any promised result lacks proof, the hook could apply to
almost any niche, or it requires fake fear, outrage, or scarcity.

## Reel package

```text
Objective:
Audience moment:
Pillar:
Format and duration hypothesis:
Selected hook / score:

Timeline
0:00–0:02 | visual | voiceover | on-screen text | retention job
0:02–0:05 | visual | voiceover | on-screen text | retention job
...continue at every meaningful edit or idea beat...
final beat | payoff | one CTA

Faceless shot list:
1. Owned or licensed source / framing / duration / edit note

Caption:
First line:
Body:
CTA:
Keyword and hashtag test:
Alt text / caption accessibility:
Music, voice, footage, and likeness rights:
Claims requiring proof:
Experiment variable:
Primary metric:
Readback checkpoint:
```

### Useful faceless modes

- product or website screen recording;
- hands-on process or studio detail;
- owned before/after evidence;
- licensed B-roll with original narration;
- animated diagrams and kinetic type;
- document, data, or workflow walkthrough;
- customer result with documented permission;
- original stills edited into a paced sequence.

Do not use anonymous stock footage that conflicts with the claim or audience.

## Carousel package

```text
Narrative:
Objective:
Audience tension:
Slide count and why every slide is needed:

Slide 1 — exact hook / visual anchor
Slide 2 — problem or stakes / evidence
Slide 3 — idea, step, or proof / visual
...
Final slide — earned summary / one CTA

Caption:
Alt text:
Source and claim notes:
Experiment variable:
Primary metric:
```

### Slide QA

- Cover identifies audience or payoff in one read.
- Each slide makes one claim or performs one step.
- Every statistic, quote, result, and comparison has a source.
- Text fits at mobile size without requiring a zoom gesture.
- Visual hierarchy, page number, and progress signals are consistent.
- Final CTA follows naturally from the delivered value.

## Stories package

Use Stories for a specific job: research, proof, objection handling, launch
support, community reply, or conversion. Do not fill a daily quota.

```text
Frame | visual | exact text | sticker/interaction | purpose | next frame
```

When using polls, questions, links, or product stickers, verify that the feature
is available on the authenticated account and that its response path is staffed.

## Voiceover markup

```text
[pace: conversational]
[emphasis: exact word]
[pause: 0.3s]
[beat]
[pronounce: word = phonetic guide]
```

Use punctuation for comprehension first. Never make the voice sound urgent,
intimate, or emotional in a way the account's real samples do not support.
    • daily-loop.md 2.8 KB 
      # Instagram Daily Loop

The daily loop produces decisions and approval-ready assets. It does not imply
automatic publishing or autonomous engagement.

## Inputs

- refreshed Account Evidence Pack;
- current content ledger and queue;
- prior post readback and due measurement checkpoints;
- active offer and conversion destination;
- audience comments, questions, DMs, search terms, and support themes the user
  is authorized to access;
- current official Instagram guidance and account status;
- lawful trend sources and cleared media library.

## Run

1. **Read yesterday.** Record post status, early measures, comments needing
   human review, and any broken link or render issue. Do not draw a performance
   conclusion before the named checkpoint.
2. **Refresh signals.** Capture only relevant audience questions, account
   insights, product events, owned proof, and current trends. Date every signal.
3. **Protect the queue.** Remove stale claims, expired links, duplicated angles,
   rights-unclear media, and posts whose context materially changed.
4. **Select candidates.** Score ideas on audience recognition, payoff, evidence,
   voice fit, and business bridge. Select 1–3 candidates with different jobs.
5. **Produce packages.** Use the Reel, carousel, Story, or static contract.
6. **QA.** Verify exact claims, source rights, identity, accessibility, CTA,
   destination, disclosure, and brand voice.
7. **Request exact approval.** Show the final copy, media list, visible identity,
   destination, timing hypothesis, and any collaborator or tag.
8. **Publish only if authorized.** Use the authenticated native surface or an
   approved API/tool permitted for the account.
9. **Read back.** Confirm permalink, identity, rendered media, caption, tags,
   collaborator state, CTA destination, and public availability.
10. **Log.** Record post ID, publish time, experiment variable, primary metric,
    measurement checkpoint, and permalink.

## Approval packet

```text
Account / visible identity:
Format:
Final media and rights source:
Final caption:
Tags / collaborators / location:
CTA and destination:
Publish time hypothesis:
Experiment variable:
Primary metric and checkpoint:
Approval status: draft | exact-content approved | published and read back
```

## Failure handling

- **No authenticated access:** return drafts and the audit worksheet.
- **Metrics missing:** label performance class `unknown` and request Insights or
  attribution evidence.
- **Composer differs from plan:** stop before publish and show the rendered
  difference.
- **Publish succeeds but readback fails:** report `publish state unknown`; do
  not retry blindly.
- **Rights or disclosure unclear:** use the halt contract and omit the asset.
- **Comments or DMs contain threats, medical/legal/financial claims, private
  data, or account-security issues:** escalate for human review; do not reply.
    • measurement.md 2.7 KB 
      # Instagram Measurement and Experiments

Measurement exists to decide what to repeat, change, or stop. Do not optimize a
proxy that is disconnected from the account objective.

## Objective map

| Objective | Primary measure candidates | Diagnostics |
|---|---|---|
| Awareness | accounts reached in target cohort | view distribution, non-follower reach, profile visits |
| Qualified growth | follows attributed to post / reach | profile visits, audience fit, unfollows when available |
| Education | saves / reach | carousel completion, average watch time, qualified questions |
| Distribution | shares / reach | sends, reposts, new-audience reach |
| Conversation | qualified comments or DMs / reach | objection themes, response burden |
| Leads | attributed leads / reach | profile actions, site taps, DM keyword starts |
| Sales | attributed sales or revenue / reach | checkout starts, lead quality, offer and margin |

The table supplies candidates, not mandatory metrics. Use only measures the
current account exposes and the user is authorized to access.

## Experiment card

```text
Question:
Comparable baseline: format / date range / n / median
Hypothesis:
One variable changed:
Constants held:
Primary metric:
Diagnostics:
Guardrail:
Minimum review checkpoint:
Decision rule: repeat | iterate | stop
Confounders:
```

Set the review checkpoint from the account's normal reach curve and publishing
cadence. Do not use a universal 24-hour, seven-day, or 10,000-view threshold.

## Readback table

```text
Post ID | format | pillar | hook family | test variable | published at
Reach | views | average watch time | saves | shares | comments | follows
Profile actions | site taps | DMs | leads | sales | source captured at
Comparable baseline | result | next decision
```

## Diagnosis rules

- High views plus low average watch time: inspect whether starts or replays
  inflate views; test the opening and message-format fit.
- Adequate retention plus low follows: inspect audience fit, profile promise,
  series continuity, and follow CTA.
- Saves without site actions: treat the post as an assist unless education is
  the primary objective.
- Shares with negative or off-target comments: inspect who is sharing and why;
  distribution is not automatically desirable.
- Site taps without leads: route to `suede-analytics` and
  `suede-site-alchemy`; do not blame the post before checking attribution and
  landing-page friction.
- Leads without sales: inspect lead quality, offer, pricing, follow-up, and
  attribution before changing content.

## Reporting

Always separate:

1. observed values;
2. computed rates and formulas;
3. comparable baseline;
4. experiment result;
5. inference and confidence;
6. the next smallest test.
  • CARD.md 5 KB 
    # Skill Card — Suede Instagram Growth

<!-- Generated by scripts/build-skill-cards.mjs — do not hand-edit. -->
<!-- Regenerate with: npm run build:cards -->

Release record for the `suede-instagram-growth` skill, following the NVIDIA skill-card template (<https://docs.nvidia.com/skills/skill-cards>). It tells a reviewer what the skill does, who owns it, what it needs, what could go wrong, and what evidence backs the release — without requiring them to open the source first.

## Description

Suede-owned Instagram growth operating system for account-specific audits, Reels, carousels, Stories, conversion mapping, calendars, and daily candidate-production loops.

Status: production. Ships in the `suede-skills` plugin (the full pack) at release 0.19.0; loads as a Claude Code / Codex agent skill from this directory's [SKILL.md](./SKILL.md).

## Owner

Jason Colapietro, Suede Labs AI (<https://github.com/JasonColapietro>). Security contact: `info@suedeai.ai` per [SECURITY.md](../../SECURITY.md).

## License / Terms of Use

MIT ([LICENSE](../../LICENSE)). The pack's combined license expression is `MIT AND BSD-3-Clause`; this skill bundles no third-party licensed material of its own.

## Use Case

Target users: developers and creators running the skill inside a Claude Code or Codex CLI session.

Use when the user names Instagram, IG, Reels, Stories, asks to analyze recent posts, grow a handle, run a daily workflow, create or repurpose Instagram content, or distinguish views from follows, leads, and sales.

Out of scope — multi-platform organic strategy (use suede-social), full video rendering or editing (use suede-video), paid Meta campaigns (use suede-ads or suede-ad-creative), analytics instrumentation (use suede-analytics), or any publish, comment, follow, like, or DM action without exact approval.

## Deployment Geography

Global. The skill is a prompt-and-script package that runs locally inside the invoking agent session; it pins no region-specific service of its own.

## Requirements / Dependencies

- A Claude Code or Codex CLI session with the `suede-skills` plugin installed (install options: <https://skills.suedeai.ai/>).
- Bundled files loaded relative to this directory: `agents/` (1 file), `references/` (4 files).
- Credentials: none are bundled or required by the skill files. Any tool or API credentials come from the host session; never paste credentials into skill files, prompts, or outputs.

## Known Risks and Mitigations

- Risk: an agent treats a quality gate as autonomous authority. Mitigation: every gate in the pack is advisory — it changes what is reported, never what the user decided; only extreme-risk findings (data loss, credential exposure, legal/rights violations, payment mistakes, irreversible public damage) pause for the user's explicit choice.
- Risk: a skill instruction is used to act outside its mandate. Mitigation: the hard limits in the skill body's "Boundaries" section, quoted below.

From "Boundaries":

- Do not publish, schedule, comment, follow, like, repost, message, or modify an Instagram account without explicit approval of the exact content and visible identity.
- Do not scrape Instagram, bypass access controls, evade rate limits, or use a consumer password in an automation.
- Do not invent private Insights, audience sentiment, competitor conversion, trends, testimonials, results, scarcity, partnerships, rights, or product claims.
- Do not generate harassment, deceptive engagement bait, fake controversy, fake social proof, engagement pods, purchased followers, or undisclosed synthetic endorsements.
- Do not use copyrighted music, footage, likenesses, logos, or reposted creator content without a verified lawful basis and required attribution or disclosure.
- Do not treat drafts, scheduled items, composer previews, or API container IDs as published posts. Require live readback.

## References

- Skill source: [`skills/suede-instagram-growth/SKILL.md`](./SKILL.md)
- Rendered reference page: <https://skills.suedeai.ai/skills/suede-instagram-growth.html>
- Security policy and reviewed scanner exceptions: [SECURITY.md](../../SECURITY.md) and [`.plugin-scanner.toml`](../../.plugin-scanner.toml) at the repo root

## Skill Output

Markdown analysis and recommendations returned in the agent's response. The skill publishes, posts, and sends nothing without the user's explicit authorization; delivery decisions stay with the user.

## Skill Version

0.19.0 — the pack is single-versioned, so every skill releases together; see [VERSION](../../VERSION) and [CITATION.cff](../../CITATION.cff) for the release identifier this card describes.

## Ethical Considerations

- The skill produces recommendations for a human decision-maker. Publishing, sending, payment, and rights decisions stay with the user.
- Its gates require verifiable claims and honest reporting; do not use the skill to fabricate claims, evidence, metrics, or attribution.
- Report suspected misuse or a security concern privately per [SECURITY.md](../../SECURITY.md); do not open a public issue for it.
  • SKILL.md 16.7 KB 
    ---
name: suede-instagram-growth
description: "Suede-owned Instagram growth operating system for account-specific audits, Reels, carousels, Stories, conversion mapping, calendars, and daily candidate-production loops. Use when the user names Instagram, IG, Reels, Stories, asks to analyze recent posts, grow a handle, run a daily workflow, create or repurpose Instagram content, or distinguish views from follows, leads, and sales. NOT FOR: multi-platform organic strategy (use suede-social), full video rendering or editing (use suede-video), paid Meta campaigns (use suede-ads or suede-ad-creative), analytics instrumentation (use suede-analytics), or any publish, comment, follow, like, or DM action without exact approval."
metadata:
  version: 1.0.0
---

# Suede Instagram Growth

Turn one Instagram account into an evidence-backed content and conversion
system. Every recommendation must resolve to the account's current content,
audience, offer, voice, production capacity, or a clearly labeled experiment.
Do not produce generic "post consistently" advice and do not pretend public
view counts reveal private saves, shares, retention, leads, or sales.

## Red Flags — Correct These First

- **"Lock this context forever."** Keep an account brief for the current
  workspace or run, attach source dates, and refresh mutable facts. Never claim
  permanent memory.
- **"Browse the last 30 posts automatically."** Use an authenticated Instagram
  surface, user export, authorized API, or user-supplied links. Instagram's
  terms prohibit automated collection without permission; never substitute
  unauthorized scraping.
- **"This went viral, so it converts."** Views are attention. Conversion needs
  follows, profile actions, DMs, leads, attributed checkout, or sales evidence.
- **"Use 15 hashtags and the best posting time."** Hashtag count, timing,
  format, and cadence are account-level tests. Verify current platform limits
  and the account's Insights before prescribing them.
- **"Run the daily workflow" means publish.** It means refresh evidence and
  prepare approval-ready candidates. Publishing is a separate authorized step.

## Operating Contract

### 1. Read existing context before asking

Read `.agents/product-marketing.md`, `.claude/product-marketing.md`, or the
legacy `product-marketing-context.md` when present. Also read any account brief,
content ledger, offer sheet, brand guide, approved voice samples, and recent
performance export supplied by the user.

Build or refresh this **Account Evidence Pack**:

```text
Handle and visible identity:
Account type: personal | creator | business | unknown
Objective: awareness | qualified followers | leads | sales | community
Primary audience:
Offer, price, and conversion path:
Voice samples and source dates:
Faceless preference and available media:
Current cadence and production capacity:
Timezone and follower-active windows:
Recent-post evidence source and coverage:
Attribution source: none | Insights | links/UTMs | CRM | checkout | mixed
Claims or topics requiring review:
Last refreshed:
Unknowns that affect confidence:
```

Start from accessible evidence. Ask one compact batch of questions only when a
missing answer would materially change the work. Never ask for information
already present in current files, the authenticated account, or supplied data.

### 2. Label every fact by evidence class

Use these labels in audits and recommendations:

- **Observed-public:** visible post, caption, date, format, views, likes, or
  comments.
- **Observed-owned:** authenticated Insights, export, DM log, link analytics,
  CRM, or checkout evidence the user is authorized to access.
- **Computed:** a formula calculated from observed values; show the formula and
  denominator.
- **Inferred:** a hypothesis to test. Never word it as an account fact.
- **Unknown:** the needed measure is unavailable. State what would resolve it.

### 3. Use the halt contract for real blockers

Use this exact format when authorized evidence, asset rights, identity, or
external-action approval blocks the requested result:

```text
HALT — <one-line blocker>
Why it blocks: <specific missing authority or evidence>
Resolve with:
1. <option>
2. <option>
3. <option, when useful>
Waiting for: <the exact item or approval>
```

Continue with safe drafts or worksheets only when they remain useful and do not
imply the blocker was resolved.

## First Action: Audit the Account

When the user supplies a handle or says to analyze the account:

1. **Verify identity and access.** Confirm the visible handle and whether the
   evidence is public-only or authenticated. Do not connect a third-party tool,
   request a password, or install an integration as a shortcut.
2. **Collect the recent cohort.** Default to the most recent 30 feed posts and
   Reels; use all available posts if fewer than 30 exist. Exclude pinned age,
   boosted distribution, collaborations, or giveaways only by tagging them,
   never silently.
3. **Collect comparable fields.** Use the schema in
   [references/account-audit.md](references/account-audit.md). Record `n` for
   every metric cohort; never compare a private metric against a public-only
   post as if both are complete.
4. **Code the content.** Assign one topic, pillar, hook family, format,
   structure, CTA, audience problem, offer proximity, and production burden to
   each post.
5. **Normalize within the account.** Compare like formats and similar
   distribution conditions. Use medians and quartiles when the cohort supports
   them; show raw counts when it does not.
6. **Map business value.** Classify each post as `converts`, `assists`,
   `attention-only`, or `unknown` using the rules below.
7. **Return the playbook.** Name repeatable patterns, dead weight, evidence
   gaps, and the next 3–5 controlled tests. Each recommendation must cite the
   post IDs or cohort behind it.

### Conversion classification

- **Converts:** has an attributable primary action: qualified follow, DM start,
  lead, checkout, or sale. Record the attribution source.
- **Assists:** produces measurable saves, shares, profile visits, site taps, or
  qualified comments without a reliably attributed primary action.
- **Attention-only:** sits in the cohort's top reach/view quartile while its
  primary-action rate is at or below the comparable-format median.
- **Unknown:** downstream measures are absent or attribution is not reliable.

Never promote `unknown` to `attention-only` or `converts` by intuition.

## Choose the Requested Mode

| User language | Execute |
|---|---|
| "analyze my account" | 30-post audit and conversion map |
| "make a calendar" | ranked 30- or 60-day test calendar |
| "create a Reel" | hook set, timed script, shot plan, caption, CTA, test card |
| "make a carousel" | selected narrative, exact slide copy, visual direction, caption |
| "repurpose this" | source atomization and a platform-native Instagram bundle |
| "analyze competitors" | lawful comparable-account pattern and whitespace audit |
| "run daily workflow" | evidence refresh and approval-ready candidate package |

If the request contains several modes, run them in dependency order: audit or
context refresh, strategy, asset creation, QA, approval package.

## Strategy and Planning

### Content pillars

Derive 3–5 starting pillars from the intersection of:

1. audience problem or desire;
2. account expertise or credible access;
3. observed response pattern;
4. offer or strategic objective;
5. repeatable production source.

For each pillar, return:

```text
Pillar:
Audience job:
Proof the account can own it:
Observed supporting posts:
Primary format hypothesis:
Business bridge:
Stop condition:
```

Do not force equal shares. Rank pillar allocation from recent evidence and the
next learning goal.

### Calendar

Build calendars as experiments, not filler. Every row must contain:

```text
Date/timezone | format | pillar | audience problem | hook | payoff | CTA
Evidence source | one variable being tested | primary metric | asset owner
Production status | approval status | readback field
```

Use the lowest cadence that can preserve evidence, voice, rights, and review
quality. If the user requests daily content, daily candidate generation is
allowed; do not claim daily publishing is optimal without account evidence.

### Competitor and trend research

Use public or authorized evidence only. Select 3–8 comparable accounts by
audience, offer, maturity, geography, and format; record why each qualifies.
Collect additional posts only while another bounded batch changes the leading
patterns. Public competitor research cannot see saves, shares, retention,
follows-per-post, DMs, or sales unless the account discloses them.

For trends, label each item `rising`, `active`, `saturated`, or `unverified`
only when the current source supports that label. Record source, observed date,
rights status, audience fit, and shelf life. A trend is optional; account fit
and source rights outrank novelty.

## Ideation and Selection

Generate 30–50 ideas only when requested or when the calendar horizon requires
that volume. Build them from the account's actual patterns:

- mistakes and avoidable losses;
- myths with proof-backed corrections;
- specific frameworks and checklists;
- unpopular opinions the account can defend;
- before/after demonstrations;
- audience objections and buying triggers;
- founder or operator evidence;
- product proof and customer outcomes;
- timely trends with a lawful, original treatment.

Score each idea from 0–20:

| Criterion | 0–4 rule |
|---|---|
| Audience recognition | 4 = target viewer identifies their problem in one read |
| Specific payoff | 4 = one concrete promised outcome, with no inflated claim |
| Evidence strength | 4 = owned proof or demonstrable source supports the idea |
| Voice fit | 4 = matches at least two supplied voice markers |
| Business bridge | 4 = natural next step connects to the stated objective |

Rank by total score, but show every component. Do not label an idea "viral."
Call it a test candidate and state why.

## Content Creation Contracts

Read [references/content-production.md](references/content-production.md) for
the full format templates.

### Reels

Return:

1. 20 hook candidates when the user asks for a hook factory; otherwise 5.
2. A selected hook with the selection score and evidence.
3. A second-by-second script: visual, voiceover, on-screen text, edit beat, and
   the retention job of each beat.
4. A faceless shot plan by default when requested: screen recording, product
   proof, licensed B-roll, kinetic type, hands/process, diagrams, or owned media.
5. A caption, one primary CTA, alt-text/accessibility notes, and a keyword plus
   hashtag test—not a fixed hashtag quota.
6. A rights and claim checklist.
7. One-variable test card and post-publication readback fields.

Choose duration from the idea and the account's comparable retention. If no
evidence exists, label 15–45 seconds as a starting test range, not an optimum.

### Carousels

Pick the narrative before writing: `problem-proof`, `mistake-fix`,
`framework`, `before-after`, `myth-evidence`, or `demo-walkthrough`. Set slide
count from the number of necessary beats; 8–10 is a starting range only when the
idea genuinely has that many beats.

The canonical slide-by-slide framework library — per-slide copy slots and a
production checklist — is owned by `suede-social`, in its carousel-frameworks
reference. Read it whenever slide-level copy slots are needed rather than
inventing a competing structure here.

Return exact text and visual direction for each slide. Slide 1 must identify
the audience tension or payoff. Every middle slide does one job. The last slide
summarizes the earned payoff and gives one next action.

### Captions and voiceover

- First line must stand alone before truncation.
- Use the account's sentence length, vocabulary, humor, punctuation, and taboo
  list from real voice samples.
- Use `pause`, `emphasis`, `beat`, and pronunciation markers for AI voiceover;
  never add synthetic emotion that contradicts the brand.
- Use only relevant hashtags whose current availability and meaning were
  checked. Do not manufacture "big/medium/niche" volume tiers without data.
- Run final copy through `suede-deslop` when it is available.

## Daily Workflow

Read [references/daily-loop.md](references/daily-loop.md), then execute:

1. Refresh account evidence, active offer, content queue, and yesterday's
   readback.
2. Review authorized trend and audience-signal sources.
3. Score candidate ideas and select 1–3 that cover distinct audience jobs.
4. Produce complete Reel, carousel, Story, or static packages.
5. Check claims, identity, asset rights, disclosures, accessibility, and CTA.
6. Return the exact approval packet; do not publish yet.
7. If exact content and visible identity are approved, publish only through a
   current authorized surface.
8. Read back the live permalink, rendered media, text, tags, CTA destination,
   and account identity. Log the post ID and measurement checkpoint.

Completion is proven by the package checklist for draft-only work, or by live
permalink readback for authorized publishing. A prepared composer is not a
published post.

## Measurement and Iteration

Read [references/measurement.md](references/measurement.md).

Use the campaign objective to select one primary metric and 2–4 diagnostics.
When denominators exist, calculate rates explicitly:

```text
save rate = saves / accounts reached
share rate = shares / accounts reached
comment rate = comments / accounts reached
follow rate = follows attributed to post / accounts reached
profile-action rate = profile actions / accounts reached
lead rate = attributed leads / accounts reached
sales rate = attributed sales / accounts reached
```

Do not combine these into a universal engagement score. Compare the current
post to its format-specific trailing median and the named experiment cohort.
Change one meaningful variable per test whenever practical.

## Suede-Owned Account Mode

When the target account belongs to Suede Labs AI, Jason Colapietro, or a named
Suede product:

- Anchor public positioning in creator ownership infrastructure,
  programmable IP, rights, provenance, registry-backed media, royalty routing,
  licensing readiness, and agent commerce. Do not reduce Suede to a generic AI
  music app.
- Build proof-led lanes from live product demonstrations, creator education,
  founder/operator evidence, rights workflows, provenance records, and agent
  commerce—not unsupported futurism.
- Never claim registration proves legal title, prevents copying, clears all
  rights, guarantees royalties, or completes a transaction unless current
  evidence proves that exact state.
- Use only `docs/assets/suede-ai-logo-transparent.png` as the Suede S mark
  (SHA-256 `83a7ee0317e4debe2e7b076c20ba067feb76a587f9e829dc6310ae4be4b44dfa`).
  Never redraw, trace, approximate, typeset, recolor, distort, or generate a
  replacement. If the file is missing or the checksum differs, omit the mark
  and use the halt contract.
- Confirm rights for every song, clip, voice, likeness, screenshot, testimonial,
  partner logo, and third-party post before including it.

## Boundaries

- Do not publish, schedule, comment, follow, like, repost, message, or modify an
  Instagram account without explicit approval of the exact content and visible
  identity.
- Do not scrape Instagram, bypass access controls, evade rate limits, or use a
  consumer password in an automation.
- Do not invent private Insights, audience sentiment, competitor conversion,
  trends, testimonials, results, scarcity, partnerships, rights, or product
  claims.
- Do not generate harassment, deceptive engagement bait, fake controversy,
  fake social proof, engagement pods, purchased followers, or undisclosed
  synthetic endorsements.
- Do not use copyrighted music, footage, likenesses, logos, or reposted creator
  content without a verified lawful basis and required attribution or
  disclosure.
- Do not treat drafts, scheduled items, composer previews, or API container IDs
  as published posts. Require live readback.

## Routing

- Use `suede-social` for cross-platform organic strategy and repurposing beyond
  Instagram.
- Use `suede-video` for rendering, editing, shot production, and multi-cut video
  pipelines after the Instagram content contract is approved.
- Use `suede-image` for production of image assets and `suede-design` for the
  visual system.
- Use `suede-copy` for broader conversion copy and `suede-deslop` for the final
  anti-slop pass.
- Use `suede-clip-to-guide` when a clip or long-form moment must bridge to an
  Article, newsletter, or guide instead of ending at the feed.
- Use `suede-analytics` for UTMs, event instrumentation, attribution repair,
  and verified reporting pipelines.
- Use `suede-ads` and `suede-ad-creative` for paid Meta campaigns.
- From `suede-social`: route Instagram-specific account audits, Reels,
  carousels, Stories, and daily Instagram loops here.

Comments (0)

Sign in to join the conversation.

No comments yet.

Reviews (0)

No reviews yet.

Related

Claude Prompt
Tighten-my-prose editing prompt
Claude Skill
Adversarial code-review skill for Claude
Claude Workflow
Deep-research workflow (fan-out + verify)
Claude Skill
Conventional-commit writer for Claude