| name | short-form-archetype-research |
| description | Route and quality-gate TikTok, Instagram Reels, and YouTube Shorts production using the local 2026-04-29 archetype research pack, blueprints, repo evidence summaries, workflow guides, and provenance rules. |
Short-Form Archetype Research
Use When
- The user asks how winning Reels, TikToks, or YouTube Shorts are made.
- You need to choose a production archetype before using
generate-short, video-render, or a lane-specific skill.
- You need repo-backed evidence, asset requirements, workflow guides,
or quality gates for a short-form format.
- You are comparing content-machine output against reference repos and
want parity checks rather than generic style advice.
Core Rule
Pick one archetype, load its blueprint, confirm required assets, and
apply the quality gates before generating or approving the short.
Do not mix archetypes casually. If the output should be hybrid, name the
primary archetype and the one secondary behavior it borrows.
Workflow
- Read routing-guide.md to choose the
archetype.
- Load the matching JSON blueprint from
assets/blueprints/.
- Load the matching workflow guide from
references/ when execution
details matter.
- Use lane-specific
examples/request.json files when a production
skill ships one.
- Check asset-pack-spec.md and
asset-inventory-and-provenance.md
before copying or generating assets.
- Apply
quality-gates-and-eval-rubric.md
before handoff.
- When the user asks for public examples, use
youtube-reference-corpus.md
and
assets/manifests/youtube-reference-corpus.json as metadata-only
references.
- When repo parity matters, read
deep-repo-extraction.md before
proposing runtime changes.
- For implementation work, read
implementation-signal-matrix.md
and
content-machine-extraction-targets.md
to map repo patterns to concrete artifacts.
- For deep implementation work, read
caption-timing-rendering-deep-dive.md,
prompt-schema-extraction.md,
render-runtime-decision-matrix.md,
and
workflow-architecture-deep-dive.md.
- For longform, crop, asset, and review implementation, read
clip-selection-and-reframing-deep-dive.md,
source-media-asset-visual-verification-deep-dive.md,
review-eval-and-governance-deep-dive.md,
and
content-machine-integration-sequence.md.
- For hook, audio, continuity, and publish implementation, read
hook-narrative-packaging-deep-dive.md,
audio-voice-music-mix-deep-dive.md,
visual-continuity-storyboard-generation-deep-dive.md,
and
platform-packaging-and-publish-deep-dive.md.
- For batch, localization, style profile, feedback, analytics, risk, and
upload-boundary implementation, read
batch-queue-gallery-production-deep-dive.md,
localization-dubbing-language-deep-dive.md,
style-template-profile-system-deep-dive.md,
feedback-analytics-learning-loop-deep-dive.md,
risk-compliance-upload-boundaries-deep-dive.md,
and
deep-implementation-backlog-sequence.md.
- For reference ingest, skill/flow/harness contracts, research evidence,
media synthesis, telemetry, provider governance, and eval datasets, read
reference-ingest-videospec-blueprint-deep-dive.md,
flow-skill-harness-contract-deep-dive.md,
research-evidence-briefing-deep-dive.md,
media-synthesis-provider-orchestration-deep-dive.md,
observability-cost-retry-cache-deep-dive.md,
and
eval-ci-golden-dataset-deep-dive.md.
- For better agent-generated SVG, HTML, CSS, and Remotion animation,
use
motion-design-coder and read
agent-motion-design-skill-systems-deep-dive.md,
remotion-svg-html-css-animation-quality-deep-dive.md,
animation-review-render-validation-deep-dive.md,
and
animation-skill-implementation-sequence.md.
Archetype Map
| Archetype | Use for | Primary references |
|---|
reddit-story-gameplay | Reddit/confession narration over gameplay or satisfying loop footage | reddit-story-gameplay.md, raga70-reddit-bot.md, dr34ming-shorts-project.md |
longform-clip-factory | Podcast, talking-head, stream, webinar, or lecture highlight extraction | longform-clip-factory.md, imgly-videoclipper.md, shortgpt.md |
topic-faceless-explainer | Topic-to-short videos from script plus stock or AI visuals | topic-faceless-explainer.md, gyoridavid-short-video-maker.md, openshorts.md |
ugc-avatar-product-short | Product, SaaS, or app promo with avatar, demo, or testimonial structure | ugc-avatar-product-short.md, claude-shorts.md |
motion-graphics-lesson | Dense educational explainers with diagrams, cards, or animated text | motion-graphics-lesson.md, samurai-ai-shorts.md |
caption-export-primitives | Caption design, subtitles, platform export profiles, and safe zones | caption-export-primitives.md, platform-snapshot.md |
Asset Rules
- Preserve provenance for copied source repo files, stock assets,
gameplay loops, screenshots, generated media, and user-supplied media.
- Treat searched YouTube, TikTok, and Reels links as metadata-only research
unless rights are separately cleared.
- Prefer the asset pack manifest in
assets/manifests/asset-pack-index.json
before searching manually.
- Use platform profiles in
assets/manifests/platform-export-profiles.json for duration, aspect,
safe-zone, and caption checks.
- Use video-asset-acquisition-policy.md
before downloading, copying, or rendering from public video references.
- If a source repo pattern depends on a missing external asset, record
the required asset class rather than pretending the template is
complete.
Output Behavior
When advising, return:
- chosen archetype and why it fits
- required source assets
- matching skill/example request or blueprint path
- repo evidence consulted
- public video references consulted, if any
- quality gates that must pass
When executing, hand off to the relevant skill or runtime script only
after the route and asset requirements are explicit.
Quality Bar
The short should match the reference repo pattern at the level of
structure, timing, caption behavior, asset provenance, and platform
constraints. Codec success alone is not enough.
Use production-blueprint-matrix.md
for expected stage artifacts and
extraction-backlog.md for known
implementation gaps.
The deeper repo pass promotes these artifact targets:
director-score.v1.jsonclip-candidates.v2.jsoncrop-plan.v1.jsonasset-ledger.v1.jsoncaption-recipe.v1.jsonvisual-match-report.v1.jsonrun-ledger.v1.jsonl
The implementation-layer pass adds:
audio-analysis.v1.jsonword-timestamps.v2.jsoncaption-groups.v1.jsoncaption-render-plan.v1.jsoncaption-export-report.v1.jsonprompt-contract.v1.jsontranscript-edit.v1.jsonrender-runtime-decision.v1.jsondecision-log.v1.jsonproject-state.v1.json
The continuation pass adds:
text-selection-candidates.v1.jsonclip-collection.v1.jsonspeaker-face-map.v1.jsonsmart-crop-track.v1.jsonsource-media-review.v1.jsonframe-samples.v1.jsonvisual-match-report.v2.jsonreview-bundle.v1.jsonprovider-selection.v1.jsoncost-log.v1.jsonarchetype-parity-fixture.v1.json
The creative-production pass adds:
hook-tests.v1.jsonnarrative-package.v1.jsonvoice-cast.v1.jsonaudio-segments.v1.jsonaudio-review.v1.jsonstyle-bible.v1.jsoncharacter-sheet.v1.jsonstoryboard-grid.v1.jsonvisual-continuity-report.v1.jsonpublish-package.v2.jsonthumbnail-plan.v1.jsondistribution-handoff.v1.json
The operations/localization/feedback pass adds:
batch-run.v1.jsonjob-queue-event.v1.jsonlsample-approval.v1.jsongallery-entry.v1.jsonbatch-output-manifest.v1.jsonbatch-policy.v1.jsonlocale-plan.v1.jsontranslation-map.v1.jsondub-script.v1.jsonvoice-transfer-plan.v1.jsonlocalized-captions.v1.jsonlocalized-publish-package.v1.jsonstyle-profile.v2.jsontemplate-contract.v1.jsonpreset-resolution-report.v1.jsonhook-library.v1.jsontemplate-safety-review.v1.jsonperformance-snapshot.v1.jsoncreator-feedback.v1.jsonvariant-outcome.v1.jsonquality-label.v1.jsonarchetype-learning-note.v1.jsonupload-boundary-review.v1.jsonrights-gate.v1.jsoncredential-requirement.v1.jsonrepost-risk-review.v1.jsonpublic-reference-use.v1.json
The agent-runtime/eval pass adds:
reference-ingest-request.v1.jsonvideo-theme.v1.jsonvideo-blueprint.v1.jsonreference-frame-analysis.v1.jsonreference-to-production-brief.v1.jsonskill-route.v1.jsonflow-execution-plan.v1.jsonharness-envelope.v1.jsonstage-input-output-map.v1.jsonagent-handoff.v1.jsonresearch-query-plan.v1.jsonevidence-pack.v1.jsonclaim-map.v1.jsonangle-selection.v1.jsonbrief-package.v1.jsonmedia-synthesis-request.v1.jsonmedia-provider-selection.v1.jsonmedia-synthesis-job.v1.jsongenerated-media-ledger.v1.jsonmedia-generation-review.v1.jsonrun-telemetry.v1.jsonlprovider-call-log.v1.jsonlretry-policy.v1.jsoncache-decision.v1.jsonbudget-summary.v1.jsongolden-archetype-dataset.v1.jsoneval-run-plan.v1.jsonarchetype-parity-report.v2.jsongolden-failure-case.v1.jsonci-eval-summary.v1.json
The agent-motion-design pass adds:
motion-brief.v1.jsonagent-motion-skill.v1.jsonmotion-code-review.v1.jsonmotion-token-map.v1.jsonremotion-motion-plan.v1.jsonsvg-motion-plan.v1.jsonhtml-css-motion-plan.v1.jsonmotion-frame-samples.v1.jsonmotion-smoothness-report.v1.jsonanimation-safe-zone-report.v1.jsonrender-performance-risk.v1.jsonmotion-review-bundle.v1.json
Use assets/manifests/runtime-artifact-roadmap.json to map these artifacts
to source signals, priorities, and Beads.
Use assets/manifests/production-quality-artifact-roadmap.json for the
clip/crop/source-review/visual-review/final-review continuation.
Use assets/manifests/creative-production-artifact-roadmap.json for
hook/audio/continuity/publish continuation work.
Use assets/manifests/operations-localization-feedback-roadmap.json for
batch/localization/style/feedback/risk continuation work.
Use assets/manifests/agent-runtime-eval-roadmap.json for reference ingest,
agent execution contracts, research evidence, media synthesis, telemetry, and
eval continuation work.
Use assets/manifests/agent-motion-design-roadmap.json for motion-design
skill, Remotion/SVG/HTML/CSS planning, smoothness review, and publish-prep
integration work.
Non-Goals
- Do not copy reference repo code into runtime without a separate
implementation task and license/provenance review.
- Do not use assets with unknown source or unclear reuse rights.
- Do not approve boxed, guttered, or caption-colliding vertical output
unless that is explicitly part of the chosen format.
