Loading
Use when the user asks to "amplify influencer content with paid media", "set up whitelisting or Spark Ads", "decide which posts to boost", "repurpose influencer content", "turn one video into multiple
The skill file stays in the publisher's repository.
Paste the copied command in a terminal. It installs this SKILL.md into that agent.
npx skills add aaron-he-zhu/aaron-marketing-skills --skill content-amplifier -a cursor -y
Use when the user asks to "amplify influencer content with paid media", "set up whitelisting or Spark Ads", "decide which posts to boost", "repurpose influencer content", "turn one video into multiple ads", or "build a UGC asset library"; produces (paid mode) a content-selection scorecard, a paid amplification strategy (whitelisting/boosting/dark posts), audience targeting, and a budget+optimization plan, or (repurpose mode) a rights-tracked content inventory, a 1-video-to-10+-asset repurposing map, per-format transformation specs, and a 30-day distribution plan. Not for gating whether a deliverable is publishable or FTC-compliant — use creator-content-auditor; not for the always-on brand posting calendar — use social-calendar-builder; not for drafting a net-new idea into platform-native packages — use social-creative-builder. 复用达人内容 / 内容放量.
Use when the user asks to "amplify influencer content with paid media", "set up whitelisting or Spark Ads", "decide which posts to boost", "repurpose influencer content", "turn one video into multiple ads", or "build a UGC asset library"; produces (paid mode) a content-selection scorecard, a paid amplification strategy (whitelisting/boosting/dark posts), audience targeting, and a budget+optimization plan, or (repurpose mode) a rights-tracked content inventory, a 1-video-to-10+-asset repurposing map, per-format transformation specs, and a 30-day distribution plan. Not for gating whether a deliverable is publishable or FTC-compliant — use creator-content-auditor; not for the always-on brand posting calendar — use social-calendar-builder; not for drafting a net-new idea into platform-native packages — use social-creative-builder. 复用达人内容 / 内容放量.
Extract more value from a frozen, approved creator asset. Two modes: paid (extend reach with paid spend — whitelisting, Spark Ads, dark posts, budget + optimization) and repurpose (reuse one asset across paid, website, email, and social — inventory, repurposing map, format specs, distribution plan). Both start from the exact version cleared by creator-content-auditor and active scoped rights. Spark/boost paths additionally require the platform state and live post they depend on; a dark post need not have been published organically.
Scope guard: this skill does NOT score a deliverable for brand alignment, message accuracy, or FTC/disclosure compliance, and it does NOT compute a STAR Trust/Appeal score or run the STAR-T1/STAR-T2 veto — that is the creator-content-auditor gate's job. This skill works the downstream lever: turning approved content into paid reach or many-channel assets, then hands off. In a product launch, this skill owns the repurposing map and the paid-amplification / distribution execution calendar (including the 30-day plan for launch content); the launch discipline's momentum-planner schedules only the launch moments and hands the distribution work here. In always-on organic social the split is the same shape: the standing brand posting calendar belongs to social-calendar-builder and net-new idea-to-multi-platform package drafting to social-creative-builder — this skill keeps repurposing of existing assets and ALL paid amplification, and the social discipline only flags boost-worthy organic winners to it.
| Mode | Use when | Core output |
|---|---|---|
| paid (default) | Extend the reach of organic creator content with paid spend | Content-selection scorecard, amplification strategy (whitelisting / boosting / dark posts), audience targeting, budget allocation, optimization playbook |
| repurpose | Reuse one approved asset across paid, website, email, and social | Rights-tracked inventory, 1-video-to-10+ repurposing map, format transformation specs, 30-day distribution plan, content library + rights tracker |
Pick with --mode paid or --mode repurpose. If unset: "boost / amplify / whitelisting / Spark Ads / dark post / paid spend / budget" → paid; "repurpose / reuse / turn one video into many / asset library / social proof on pages / multi-channel rollout" → repurpose. If the request spans both (e.g. "cut ad variations and plan the paid spend"), run repurpose first to produce the assets, then hand to paid — do not silently merge; state which mode you ran.
Shortest invocation:
Which frozen auditor-approved asset with current active paid rights should we amplify from [campaign]? # paid
How can we repurpose this frozen approved asset within its evidenced active destination scope? # repurpose
Common scenarios:
--mode paid: Create a $5,000 TikTok/Instagram plan from these frozen approved assets and supplied current active scoped-rights records
--mode repurpose: Build a 30-day plan from these 3 frozen approved videos and supplied current active rights records covering the requested destinations
Output expectation — paid: every evidence-complete candidate scored, tiered, and given a spend that sums to budget, plus a scale/pause playbook. Any candidate missing organic-performance evidence or a Hook/Message/Quality/CTA observation remains NOT_SCORED/NEEDS_INPUT with no /25, rank, tier, or spend. repurpose: every source asset rights-tagged, at least one eligible asset mapped to 3+ formats across 2+ channels, plus a dated distribution plan.
creator_ref values plus transient creator locators, platform, content type, dated organic reach/engagement/views evidence, separate dated Hook/Message/Quality/CTA observations, amplification budget, campaign objective (awareness/traffic/conversions), target platforms, any prior performance data the user provides, and each candidate asset's current rights status/evidence plus existing placements when present.creator_ref values plus transient creator locators and platforms, exact frozen approved_asset_ref, opaque authorized source_ref, usage rights per asset, rights status with observation time/evidence, original performance metrics, target channels, and existing placements. For atomizing a source, the pasted transcript/caption/review text is transient evidence bound to that three-ref source identity; a raw handle or content URL never becomes source_ref.memory/hot-cache.md when memory-management is active.memory/influencer/content-amplifier/YYYY-MM-DD-<topic>.md only with exact WARM-save authorization. Saved artifacts and handoffs reuse an explicitly carried opaque creator_ref, or a verified creator-registry aggregate ID; otherwise generate one random creator-<UUIDv4> once for the lineage. Never persist a raw handle, creator name, profile/content URL, email, provider ID, or deterministic hash as identity. Keep only creator_ref, frozen asset/approval refs, opaque evidence/rights/contact/placement refs, and non-identifying campaign data; every saved atom source must carry creator_ref + exact approved_asset_ref + opaque source_ref. If no authorized resolver exists, mark identity unresolved and require the transient locator again when needed.memory/hot-cache.md. A newly observed revocation, dispute, expiry, or other rights-status change is not made canonical here: offer only a separately authorized handoff to the existing creator-registry proposal path with the exact status evidence. This skill does not append or accept that proposal.Emit the standard shape from skill-contract.md §Handoff Summary Format. State which mode ran. Label every metric Measured / User-provided / Calculated / Estimated / Unknown. Never fill a missing CPM, ROAS, view count, rights date, or decision threshold with an example value: ask for the export or use
Unknown/NEEDS_INPUT. Organic performance and the Hook/Message/Quality/CTA observations are selection evidence, not forecast fields: if any is missing, do not substituteEstimated; set that asset toNOT_SCORED/NEEDS_INPUTand emit no/25, rank, tier, or spend. Use Estimated elsewhere only when the user supplied or approved the calculation inputs and assumptions, and show that basis.
This family is Tier 1: both modes work with no live integrations. Ask the user for the mode's inputs and produce the supported portion of the artifact from those. Never invent reach, engagement, CPM, ROAS, rights numbers, or selection observations. Missing organic reach/engagement/views evidence or Hook/Message/Quality/CTA evidence is never Estimated; it makes that asset NOT_SCORED/NEEDS_INPUT. Other forecast fields may be Estimated only from supplied or explicitly approved inputs and assumptions; otherwise use Unknown/NEEDS_INPUT.
Where a connector could sharpen the output (all optional, opt-in Tier 2/3):
~~social platform analytics — pull organic reach, engagement rate, and view counts (both modes) instead of asking the user to paste them.~~ad platform (Meta Ads Manager, TikTok Ads Manager, Google Ads) — read live CPM/CTR/CPC/ROAS for the paid optimization playbook, and confirm Spark Ads / Partnership Ad authorization status.~~influencer database — verify creator audience demographics for lookalike targeting (paid); transiently resolve creator locators while retaining only creator_ref, platform, and contract-rights refs in saved repurpose artifacts.~~DAM / asset library — store and tag processed assets; enforce the naming convention (repurpose).~~CRM — supply retargeting/exclusion audiences (paid); reconcile creator records with usage-rights expirations (repurpose).See CONNECTORS.md for the verified free/keyless recipe per category. None are required; absent a connector, the user supplies the numbers.
Select the mode first (see Mode selector), then run that mode's steps. Each step has a fill-in template in references/templates.md — produce the populated artifact, do not skip the table.
creator_ref, platform/type, and dated organic reach/ER/views values with evidence refs. Do not create an organic-performance score from an absent, undated, or unevidenced metric. Paid Step 1 template.source_ref and observed_at. Only when all five are present may the asset receive /25, rank, tier, or spend. If any component or its evidence is missing, preserve supplied observations, set score_state: NOT_SCORED and execution NEEDS_INPUT, list the exact gaps, and do not estimate, normalize, prorate, or hand-calculate a partial total. Paid Step 2 template.NEEDS_INPUT and no hypothetical allocation. Paid Step 5 template.Return the populated artifact inline. Offer its exact WARM path for save authorization, then ask separately before any HOT promotion. Any ad-account setup, upload, launch, or spend is a further external mutation requiring exact action approval.
status_observed_at, and status_evidence_ref. Repurpose Step 1 template.[campaign]_[creator_ref]_[platform]_[type]_[variation]_[date] naming convention, and metadata fields. Repurpose Step 6 structure.status: active with status_observed_at and status_evidence_ref, then verify that the grant is unexpired and covers that channel, territory, format, and use. , , , , and out-of-scope rights remain blocked for new use. Put affected existing placements in the queue with destination, owner, , and ; a blank completion reference means removal is not proven. .For slicing one source into many output atoms, apply the 7-tier extraction in references/atom-extraction.md. Bind every atom to creator_ref, exact approved_asset_ref, and an opaque authorized source_ref. Any atom ranking, virality score, near-duplicate threshold, paid-placement choice, or hero-placement choice requires a user-approved rule or source-dated rule ref; without one, retain the atoms as NOT_SCORED/NEEDS_INPUT and do not auto-rank/drop/select them. Return the populated artifact inline; request one exact authorization for its WARM save and a separate authorization for any HOT promotion.
For either mode, the removal queue is a hand-operated follow-up view, not authorization to remove or edit anything. Never pause an ad, delete a post, unpublish a page, edit a platform, or mark completion_ref without separate exact action approval and completion evidence. When rights status changes, offer the exact evidence to the existing creator-registry proposal workflow only after separate authorization; do not create a new rights registry or treat the WARM tracker as canonical.
NOT_SCORED/NEEDS_INPUT with no total/rank/tier/spend; (3) either mode lacks the frozen asset's auditor-approval reference, has assets whose rights are expired, revoked, disputed, unknown, lack dated status evidence, or do not cover the intended channel/territory/format/use — return NEEDS_INPUT for the exact approval or rights gap before recommending activation or reuse, because neither approval nor a restricted grant may be guessed through.paid — User: "The dated organic export supplies views/ER and Hook observations for five frozen, auditor-approved assets with active scoped paid rights, but we have no Message, Quality, or CTA observations. Allocate our $5,000 budget."
Evidence basis: organic metrics, Hook observations, frozen approval refs, and rights status/scope are user-provided; Message, Quality, and CTA evidence is missing.
| Creator Ref | Organic | Hook | Message | Quality | CTA | Score State | Total / Rank / Spend |
|-------------|---------|------|---------|---------|-----|-------------|----------------------|
| [creator_ref per asset] | supplied | supplied | Unknown | Unknown | Unknown | `NOT_SCORED/NEEDS_INPUT` | none |
No `/25`, rank, tier, or allocation is emitted. Request dated Message/Quality/CTA observations with opaque evidence refs; never fill them with `Estimated` or force a partial total.
repurpose — User: "Use three supplied frozen, auditor-approved assets mapped to creator_ref-1, creator_ref-2, and creator_ref-3. Their rights records are active, dated, evidenced, unexpired, and cover the requested derivative edits and US paid, website, email, organic-social, and YouTube destinations. Build a 30-day reuse plan." → Map only those evidenced uses and retain creator_ref plus opaque asset/evidence refs in the plan.
If either request omitted the frozen approval reference or current scoped-rights evidence, return NEEDS_INPUT with those exact gaps and do not rank, allocate, transform, or schedule the asset. Paid selection additionally stays NOT_SCORED/NEEDS_INPUT when organic metrics/evidence or any Hook/Message/Quality/CTA observation/evidence is missing.
Full rankings, strategies, setups, and both worked examples: references/templates.md.
STAR-T1 FTC disclosure, STAR-T2 claim integrity) that creator-content-auditor enforces before this skill runs.After delivering findings inline, ask: "Save these results for future sessions?" If yes, obtain exact authorization for memory/influencer/content-amplifier/YYYY-MM-DD-<topic>.md and write the one-line verdict/headline, top 3-5 actionable items, open loops or blockers, and opaque source references. Saved identity is creator_ref only; remove raw handles, names, profile/content URLs, emails, and provider IDs. Saving does not authorize HOT promotion or any ad-platform mutation. This skill hands veto-like risks (missing disclosure, unsubstantiated claims) to creator-content-auditor rather than judging them here.
Primary:
Alternates:
Termination: maintain a visited-set this session. If a recommended target (including the sibling mode of this skill) already ran, STOP and report the chain complete rather than re-invoking it. Max chain depth 3. When routing is ambiguous, present the options and stop instead of auto-following.
NOT_SCORED/NEEDS_INPUT with no total/rank/tier/spend; (2) a budget allocation across only evidence-complete eligible content, objectives, and platforms sums to the stated budget; (3) an optimization plan with supplied KPI targets and scale/pause rules is recorded, or those rules stay NEEDS_INPUT; (4) every selected asset has current active rights evidence covering the intended use.active | expired | revoked | disputed | unknown status, status_observed_at, and status_evidence_ref; (2) at least one source asset is mapped to 3+ distinct output formats across 2+ channels; (3) a dated distribution plan with an asset checklist exists; (4) any existing placement affected by a non-active or out-of-scope status appears in the manual removal queue.active with a dated evidence reference; separately confirm the grant is unexpired and scoped to the intended platform, territory, format, and paid use. expired, revoked, disputed, or unknown fails closed for new use. Spark/boost or other existing-post methods also require the matching live post and platform authorization; dark-post setup does not require prior organic publication. Missing approval, required live state, or usable rights stops activation with NEEDS_INPUT; none may be inferred. Paid Step 7 guide.expiredrevokeddisputedunknowndue_atcompletion_ref