2 added, 2 removed. Audit A to A.
---
name: capability-spotlight-video
description: Plan and submit a recurring series of diverse, premium Ace Data Cloud sales films. Each run discovers live capabilities, proves a real buyer outcome, assembles a rich provenance-bound material set, chooses a structurally distinct creative genome, and routes a freeform Pro production through Maestro general-video.
license: Apache-2.0
metadata:
author: acedatacloud
version: "5.0"
connections: [acedatacloud/acedatacloud]
compatibility: Requires the AceDataCloud connector, Maestro MCP, and only the capability MCPs authorized for the selected campaign.
---
# Ace Data Cloud Campaign Film Director
Create one **premium, product-specific sales film** per production run. The series must vary in subject, visual world, film structure, camera language, rhythm, transitions, sound, voice, and material mix—not merely palette or labels.
Read first:
- [brand kit](references/brand-kit.md)
- [topic registry](references/topic-registry.md)
## Phase gate — preview means planning only
When the caller says `preview-only`, this rule overrides every later production section:
- load only this Skill and the explicitly allowed AceDataCloud discovery MCP;
- query catalog/docs/OpenAPI/pricing only;
- never attempt to load Maestro or any capability/media-generation MCP;
- never generate media, call `publish_artifact`, or start a production task;
- output the three candidates, selected Blueprint, Creative History marker, completion marker, and a **text-only future handoff contract**, then stop.
The production sections below describe what a separately authorized future task must do. They are not actions for the preview run.
## 1. Discover live truth
Load the AceDataCloud MCP and verify candidates from live public sources:
1. `acedatacloud_list_services(private=false, limit=300)` and `acedatacloud_get_service`;
2. `acedatacloud_list_apis` and `acedatacloud_get_api_spec`;
3. `acedatacloud_search_docs` and `acedatacloud_get_doc`;
4. `acedatacloud_list_model_catalog` when models matter.
A candidate is eligible only when the public service/API/docs are live and this run can execute or reuse every material required for a convincing film. Never expose private routing/source details, private IDs, signed URLs, account data, or credentials.
## 2. Generate candidates and maximize creative distance
Read `{{run_count}}`, `{{date_iso}}`, `{{last_output}}`, `{{creative_policy}}`, and `{{format}}`. Call `maestro_list_tasks(limit=30)` and inspect recent `ADC-SPOTLIGHT:v1` artifact summaries when available.
Build at least one eligible candidate in every mode before choosing:
1. **Single Capability** — one product/model, one distinctive buyer outcome;
2. **Workflow Campaign** — 2–4 services producing one visible outcome;
3. **Platform Story** — platform value proved through real cross-modal results or product lifecycles.
For each candidate write a complete Creative Genome:
`CAMPAIGN_MODE FILM_ARCHETYPE VISUAL_WORLD SHOT_GRAMMAR RHYTHM_PATTERN OPENING_DEVICE CLIMAX_DEVICE TRANSITION_SYSTEM BRAND_BEHAVIOR VOICE_ID SOUND_WORLD PALETTE_FAMILY TYPE_SYSTEM MATERIAL_MIX`
Score live truth, sales specificity, hero strength, authorization, price proof, material richness, and **creative distance**. Apply these exclusions from the last 12 artifacts:
- no recent 8 `TOPIC_ID`, buyer job, or `HERO_CASE_ID` repeat;
- no recent 6 `FILM_ARCHETYPE` or `OPENING_DEVICE` repeat;
- no recent 5 visual-world or shot-grammar repeat;
- no recent 4 campaign-mode, climax-device, or rhythm-pattern repeat;
- no recent 3 transition-system, voice-family, sound-world, or palette-family repeat;
- a returning service must change at least four of buyer, job, hero, archetype, visual world, and offer.
The scheduled task supplies the **mandatory lane**, `PRIMARY_FOCUS`, and `SECONDARY_FOCUS` as server-authoritative values. Never recompute or override them. At least one candidate must be eligible for that mandatory lane, and the selected candidate must come from it unless live truth/material evidence makes the whole lane ineligible:
1. image craft / typography / edit fidelity;
2. cross-service workflow transformation;
3. video / motion / reference control;
4. Agent / MCP / artifact / scheduled lifecycle;
5. audio / music / voice with an audio-reactive world;
6. search / extraction / structured-data outcome;
7. platform brand anthem with three real surfaces/modalities;
8. deployment / bot / operational lifecycle with sanitized real status evidence.
The mandatory lane is a hard selection lock, not a score bonus. Candidate A/B/C labels are not a global ranking: first identify the lane candidate, then compare creative treatments inside that lane. A higher score from another lane cannot override it.
Before final output, run this mechanical lane preflight:
- write `MANDATORY_LANE=<1-8>:<lane-name> | PRIMARY_FOCUS=<id> | SECONDARY_FOCUS=<id>` immediately after `CREATIVE-HISTORY:v1`, before long Creative Genome fields;
- if the selected topic belongs to it, write `LANE_RESULT=selected`;
- otherwise the entire lane must be ineligible, write `LANE_RESULT=blocked` and `LANE_BLOCKED=<specific live-truth/material reason>`, then select the least-recent eligible family;
- never emit a selected topic from another lane with `LANE_RESULT=selected`.
The completion marker repeats `lane=<1-8> | lane_result=selected|blocked`. Never silently collapse back to image. Recent-history exclusions still apply inside each lane.
Use `date_iso:run_count` only as a deterministic tiebreaker; do not use unconstrained randomness. `creative_policy` influences scoring but never becomes a fixed shot template:
- `auto-diverse`: maximize distance from recent work;
- `brand-anthem`: prefer platform desire and brand-led payoff;
- `product-cinematic`: prefer one product and sensorial/cinematic proof;
- `workflow-story`: prefer a visible multi-step transformation.
Stop rather than fall back to a generic showcase when history, authorization, real hero evidence, pricing, or rich materials are unavailable.
## 3. Prove workflow edges and real outcomes
For workflow candidates, build a runtime capability graph before generation.
**Node contract:** service/model, input/output modality, sync/async lifecycle, poll API, reference roles, authorized MCP, price payload, and accepted asset.
**Edge contract:** every output must legally become the next input, proved by OpenAPI or an executed request. Never combine services because their names sound compatible. A workflow uses 2–4 services and sells one outcome, not a logo parade. A platform story requires at least three real results or product surfaces across three modalities/lifecycles; catalog cards alone are forbidden.
Load only the selected 2–4 MCP servers from the authorized pool. A capability outside that pool is eligible only when a public accepted artifact already exists with explicit provenance.
## 4. Reuse tasks before spending
For generated hero media, the first capability-specific call MUST be the provider's `list_tasks`/batch list with a 24-hour `created_at_min` window or closest equivalent. Do not generate before this lookup.
- reuse a matching completed request and accepted public URL;
- resume a matching pending task by ID;
- create only when no match exists or the match terminal-failed;
- with no list tool, resume only a `SOURCE_TASK` from this task's `last_output`.
Poll according to the returned interval until terminal. For activation/source preparation, reserve up to two minutes (for example eight 15-second polls). If still pending, record one `ADC-SPOTLIGHT-SOURCE:v1` draft artifact with task ID and fingerprint, make no completion claim, and end without Maestro.
Inspect every selected asset at full resolution. Before Blueprint approval, perform a complete body download for every selected URL, require non-zero bytes, decode at full resolution, and inspect pixel-level semantic fit to `HERO_CASE` and its assigned material role. A metadata, filename, URL, or task status alone never proves availability, visual content, or semantic relevance. A failed download, decode, or semantic-fit check makes that asset ineligible: choose another asset or candidate and do not submit Maestro. Generated media MUST use a real accepted result. Product/Agent/deployment stories use faithful, sanitized screenshots or lifecycle evidence. Full request JSON is evidence-only; customer-facing frames never contain internal review or provenance language.
## 5. Require a rich material plan
Record provenance privately, but never place labels such as `accepted output`, `decoded proof`, `evidence bundle`, task hash, or review gate in the film.
### Single Capability
Require at least four useful visual roles: real input/before, accepted hero, two real details/alternates/decoded derivatives, and optionally product context/UI/brand environment. One unchanged still may not be stretched through the film. If the material cannot support varied compositions, choose another candidate.
### Workflow Campaign
Require 5–8 roles covering 2–4 real stages. Every stage has real input/output/task evidence and pushes the same final outcome. The final result is the hero.
### Platform Story
Require at least three distinct modalities or product surfaces and three real results. At least one role must contain real motion or a real UI/task lifecycle. A catalog-only slideshow is forbidden.
### Audio, Agent, and Deployment
- audio: real audio/result, waveform or timing data, cover/lyrics/context, and an audio-reactive visual world;
- Agent: real tool trace, delivered artifact/result, and memory/schedule lifecycle—never fake chat bubbles;
- deployment: real sanitized configuration/status/lifecycle screens; without them the candidate is ineligible.
## 6. Write Sales Blueprint v3
Before Maestro, write:
```text
ACE-DATA-CLOUD-SALES-BLUEPRINT:v3
CAMPAIGN_MODE: single | workflow | platform
PRODUCTS: <1–4 services/models>
FINAL_OUTCOME: <one buyer result>
AUDIENCE: <specific buyer/context>
PAIN: <concrete frustration>
PROMISE: <the outcome being sold>
HERO_CASE: <real representative result>
UNIQUE_ADVANTAGE: <visible differentiator>
WORKFLOW: <ordered proven edges, or N/A>
PRICE: <payload-bound live proof>
OFFER: <why start now>
CTA: <action + destination>
FILM_ARCHETYPE / VISUAL_WORLD / SHOT_GRAMMAR / RHYTHM_PATTERN
OPENING_DEVICE / CLIMAX_DEVICE / TRANSITION_SYSTEM
MATERIAL_MIX: <role→URL map with private provenance>
BRAND_BEHAVIOR / VOICE / SOUND_WORLD / COLOR_LOGIC / TYPE_SYSTEM
FORBIDDEN_REPETITIONS / QUALITY_GATES
```
Do **not** write a fixed six-scene or fixed timestamp storyboard. Describe emotional beats, hero moments, material roles, and 3–6 measurable quality gates; the general-video Director owns scene count, timing, layouts, and transitions.
The Blueprint must still answer: who buys, what changes, why this product is distinctive, which real result proves it, what it costs, and what to do next. Narration length follows the selected format and must finish before the CTA hold.
Immediately after Blueprint v3, write:
```text
ACE-DATA-CLOUD-CONTENT-COVERAGE:v1
PRIMARY_FOCUS: <server-authoritative id>
SECONDARY_FOCUS: <server-authoritative id>
BUYER_PAIN: <specific buyer and visible current-state pain>
PROMISED_EFFECT: <observable buyer result>
PROOF_ASSET: <real result that substantiates the effect>
INTEGRATION_ANCHOR: <exact public endpoint plus action/lifecycle>
VALUE_ANCHOR: <payload-bound Credits plus concrete outcome purchased>
MECHANISM: <why this input becomes this output>
CTA: <action plus destination>
SEMANTIC_BEATS: <flexible content beats, not a fixed timestamp or scene template>
```
Every film must communicate buyer/pain, promise/effect, real proof, one exact integration anchor, one payload-bound value anchor, and CTA through caption, narration, or unmistakable final-byte visual evidence. The primary focus receives one complete explanatory beat; the secondary focus receives one supporting beat; remaining anchors stay concise. Reject a candidate when this cannot fit legibly in the selected duration. Do not restore a fixed timestamp storyboard.
The five focus treatments are:
- `effect-proof`: establish a relevant current state, show the accepted result, state the visible change, and connect it to the buyer outcome. A beauty shot without before/change/outcome fails. It may compress but never omit integration and value anchors.
- `api-integration`: show the exact public endpoint, a 4–7-line safe request using `$ACEDATACLOUD_API_KEY`, input/action, sync/async lifecycle and polling/retrieval when applicable, and the real terminal result. Explain how the request produces the buyer outcome; never show a raw provenance dump.
- - `price-value`: show the payload/model/size/duration/count/reference dimensions that determine live payload-bound Credits and the concrete outcome purchased. A detached price card, unqualified “from” price, or unproved fiat/savings claim fails.
+ - `price-value`: show the payload/model/size/duration/count/reference dimensions that determine live payload-bound Credits and the concrete outcome purchased in the same composition for at least 2 continuous seconds. A detached price card, unqualified “from” price, or unproved fiat/savings claim fails. Final review must supply targeted final-byte frames at the start, midpoint, and end of that continuous range; Blueprint text, narration alone, or an artifact summary cannot prove this focus.
- `workflow-mechanism`: show ordered input→operation→result, legal proved edges and lifecycle. A single-service film explains real controls rather than inventing a workflow; a multi-service film explains why every stage is necessary and sells one outcome.
- `buyer-transformation`: name the buyer/context, recognizable pain, changed working state, real proof, distinctive advantage, and a specific reason to act. Generic “work smarter” language fails.
Palette, voice, format, or archetype changes cannot satisfy a content focus. Focus is a hard treatment lock inside the mandatory lane and does not replace existing Creative Genome distance rules.
## 7. Choose a professional, product-derived creative world
The registry is an anchor library—examples, not an allowlist. Derive a concrete world from the product's meaning, not a generic cyan dashboard.
Film archetype vocabulary includes brand anthem, product reveal, workflow transformation, launch trailer, kinetic manifesto, cinematic UI demo, before/after case film, multi-modal montage, metric-to-human payoff, and visual poem.
Shot grammar includes macro→wide, match cuts, continuous camera, split-screen convergence, kinetic-type slam, parallax editorial, UI depth flythrough, and object-led transitions.
Rhythm includes cold-open burst, fast-fast-slow-climax, crescendo montage, tension→release, precision pulse, and luxury restraint.
Preserve source-media palettes. Ace identity remains canonical, but its **behavior varies**: energy rail, mask, registration system, spatial frame, light, typographic cadence, or audio sting. Do not default to the same corner watermark, cyan UI frame, centered title card, evidence card, or proof checklist.
## 8. Route production through Maestro general-video
The Maestro production prompt begins with the non-routing line:
`ACE-DATA-CLOUD-BRAND-FILM:v1`
It MUST NOT contain `ADC-SPOTLIGHT:v1`; that marker is reserved for the outer artifact summary. Explicitly require:
- `scenario=auto`, `quality=pro`, and route to `/general-video`;
- never use `spotlight_prepare.py`, `spotlight_renderer.py`, or the fixed six-scene Spotlight template;
- read general-video's `house-style.md` and `video-composition.md`;
- write a one-sentence concept angle, embedded font pairing, and foreground-density plan;
- run prompt expansion for multi-scene work;
- name the rhythm pattern and build hero layouts before animation;
- use modular sub-compositions for 3+ hard cuts and intentional transition recipes;
- use Fish narration timing and a product-appropriate real music/SFX world;
- honor the Blueprint's Creative Genome and material-role map without displaying private provenance.
Formal production uses **Pro**, never Standard as a quality proxy. Choose format from `{{format}}`; `auto-diverse` must also avoid the recent three formats and fit the content/channel.
Treat Pro turns as a delivery budget, not an invitation to explore: finish the initial render by turn 140; use only `check` plus a contact sheet and at most six targeted frames for each review; Do not run deprecated `validate` or `inspect`. The visual-review call MUST NOT launch an Agent or recursively inventory evidence—read the manifest, contact sheet, and targeted frames directly. Consolidate all blockers into one repair, run one final `check`, one final render, and one final-byte review. If that review exposes a new release blocker, allow one blocker-only contingency patch followed by one re-render and one confirmation review; do not change story, audio, assets, or already-passing scenes. A second contingency loop is forbidden. Then confirm delivery by turn 240 and stop optional inspection when any milestone is at risk.
Immediately after Maestro accepts the unique UUID task, call `publish_artifact` before waiting or polling. Exact-ID `task_already_exists` is a successful idempotent replay: do not generate another UUID or resubmit.
The artifact summary—not the Maestro prompt—begins:
`ADC-SPOTLIGHT:v1 | BLUEPRINT=v3 | PRIMARY_FOCUS=<id> | SECONDARY_FOCUS=<id> | COVERAGE_CONTRACT=v1 | FAMILY_ID=<id> | TOPIC_ID=<id> | DEMO_ID=<id> | WORKFLOW_ID=<id-or-na> | FILM_ARCHETYPE=<id> | VISUAL_WORLD=<id> | SHOT_GRAMMAR=<id> | RHYTHM_PATTERN=<id> | OPENING_DEVICE=<id> | CLIMAX_DEVICE=<id> | TRANSITION_SYSTEM=<id> | STYLE_ID=<id> | PALETTE_ID=<id> | VOICE_ID=<id> | SOUND_WORLD=<id> | FORMAT=<id> | ASSET_HASHES=<hashes> | EXECUTED_APIS=<paths> | EVIDENCE_URLS=<urls> | MAESTRO_TASK=<uuid> | SUBMISSION=accepted`
Follow it with the complete Blueprint and role→asset provenance. The outer task ends after `recorded=true`; it must not claim the asynchronous film is complete.
## 9. Professional review contract
The Blueprint defines archetype-specific gates plus these invariants:
- decoded frame 0 contains clear brand or product desire—not blank/loading/disclaimer;
- for a 30-second format, final-byte `ffprobe` duration is 29.95–30.05 seconds; any other requested format uses the same ±0.05-second tolerance;
- the selected real hero occupies a visually primary region for at least 40% of runtime, including one near-full-frame hold of at least 3 continuous seconds; authored type/UI may support it but never substitute for it;
- composition variety and pacing match this run's genome;
- no static material is dragged beyond the limit appropriate to the archetype;
- customer copy contains no internal review language, private sourcing facts or invented claims;
- narration/audio are complete and the CTA is readable.
For both initial and final-byte review, create this matrix from the rendered MP4—not from planning prose:
```text
FINAL-BYTE-COVERAGE:v1
| Dimension | Treatment | Status | Timestamp/range | Caption | Narration | Visual/result |
| buyer-pain | base/primary/secondary | pass|na|fail | ... | ... | ... | ... |
| effect-proof | base/primary/secondary | pass|na|fail | ... | ... | ... | ... |
| api-integration | base/primary/secondary | pass|na|fail | ... | ... | ... | ... |
| price-value | base/primary/secondary | pass|na|fail | ... | ... | ... | ... |
| workflow-mechanism | base/primary/secondary | pass|na|fail | ... | ... | ... | ... |
| CTA | base | pass|fail | ... | ... | ... | ... |
```
- A `pass` needs a final-MP4 timestamp/range plus caption, narration (or explicit `none`), and visible result evidence where applicable. `na` may waive only a genuinely inapplicable focus enhancement; it cannot waive base integration or value anchors. Provenance or Blueprint text that never reached the MP4 does not count. Base sales coverage, primary and secondary focus depth, and CTA must all pass; contradictions between captions, narration, API/price truth, and visible results are blockers.
+ A `pass` needs a final-MP4 timestamp/range plus caption, narration (or explicit `none`), and visible result evidence where applicable. A primary `price-value` pass additionally needs one ≥2-second range and start/mid/end decoded frames where payload, Credits, and purchased outcome are simultaneously legible. `na` may waive only a genuinely inapplicable focus enhancement; it cannot waive base integration or value anchors. Provenance or Blueprint text that never reached the MP4 does not count. Base sales coverage, primary and secondary focus depth, and CTA must all pass; contradictions between captions, narration, API/price truth, and visible results are blockers.
Make an actual `Skill` call with `skill="visual-review"` against complete initial evidence. Fix blockers in one concentrated pass. Rebuild evidence from the exact final MP4 bytes, measure duration with `ffprobe`, verify hero runtime and the coverage matrix from the final contact sheet/transcript, and make a second actual confirmation call even when no initial blocker exists. If duration, hero-share, audio, CTA, final-byte coverage, primary/secondary focus depth, or either review fails, Maestro MUST NOT return an accepted submission. The Reviewer receives the Blueprint, Content Coverage contract, Creative Genome, role map, and evidence and judges both sales clarity and **its chosen director language**, not whether it resembles previous episodes.