AI Skill Report Card
AI Automation Overlay for Data To Product Repackaging
AI-Automation Overlay for Data-to-Product Repackaging
Quick Start10 / 15
Reframe every atomic block from the original 37-block pipeline as a typed function: block(input_schema, confidence_threshold) → {output_schema, confidence, logs}. Chain these into a DAG orchestrated by a planner agent. Route low-confidence outputs to human review. This skill converts a manual repackaging workflow into an agentic automation pipeline.
classify_source(file) → {source_type, modality, structure_schema, parser_route, confidence}
if confidence < 0.8 → escalate_to_human()
else → route_to_parser(parser_route)
Recommendation▾
Missing YAML frontmatter entirely (no 'name' or 'description' fields) — this is a critical format violation that should be added immediately for the skill to be discoverable/loadable.
Workflow13 / 15
Progress:
- [ ] Step 1: Define orchestrator + shared memory (vector DB, metadata store, asset store, job queue)
- [ ] Step 2: Map each atomic block to an agent/function with I/O schema
- [ ] Step 3: Assign AI capability + tooling per block
- [ ] Step 4: Set confidence thresholds and HITL gates
- [ ] Step 5: Wire blocks into DAG by phase dependency
- [ ] Step 6: Execute, log, and route exceptions
- [ ] Step 7: Close feedback loop into re-triggering upstream blocks
Phase 0 — Orchestration Layer (build this first)
- Shared memory: vector DB (embeddings), metadata store, asset store, version control, job queue.
- Agent roles: Source Scout, Rights Auditor, Quality Scorer, Atomizer, Context Reframer, Metadata Tagger, Transformer, Assembler, Renderer, QA Agent, Publisher, Valuator.
- Execution pattern: event-driven DAG; every node =
input_schema → function → output_schema + confidence + logs. - Escalation rule: confidence < threshold, legal ambiguity, brand risk, or high print-run cost → human gate.
Phase 1 — Source Acquisition (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 1.1 Source Type Classification | classify_source(file) → {type, modality, schema, route} | VLM/LLM + file sniffer + OCR | confidence < 0.8 or unknown format |
| 1.2 Rights & Licensing | audit_rights(source) → {status, license, allowed_derivatives, risk_flags} | Legal RAG + license classifier | any risk_flag = true |
| 1.3 Source Quality | score_source(source) → {completeness, formatting, metadata_density, fixes} | Heuristics + LLM-as-judge | score below threshold |
| 1.4 Source Aggregation | aggregate(targets) → {corpus, dedupe_map, embeddings, manifest} | Crawlers/connectors + MinHash + embeddings | scope/ToS approval |
Phase 2 — Atomization & Extraction (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 2.1 Idea Identification | extract_atoms(text) → [{type, text, location, embedding}] | LLM structured extraction, semantic chunker | precision/recall spot-check |
| 2.2 Unit Qualification | qualify_atom(atom) → {standalone_score 1-5, reason} | LLM-as-judge / reward model | score ≤ 3 |
| 2.3 Context Repatriation | repatriate(atom) → {atom+context, entailment_score} | LLM generation + NLI verifier | low entailment or high-stakes domain |
| 2.4 Element Metadata Tagging | tag_element(element) → {type_tags, confidence} | Zero-shot classifier / ontology tagger | mis-tag correction |
| 2.5 Format Inventory | inventory(assets) → {format_matrix, missing, conversion_tasks} | File-type detector + code interpreter | conversion priority approval |
Phase 3 — Transformation & Enrichment (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 3.1 Pretextual Wrapper | generate_wrapper(source, atoms) → {intro, commentary, theme} | LLM + RAG | editorial voice review |
| 3.2 Cross-Platform Adaptation | adapt_channels(atom) → {email, web, social, sms, print} | LLM + constraint engine | brand/compliance review |
| 3.3 Format Translation | translate_format(text) → {faq, list, table, qa} | LLM structured output | fidelity check |
| 3.4 Style & Branding | apply_brand(content, kit) → {styled_assets} | LLM + image gen + CSS engine | brand approval |
| 3.5 AI Script Enrichment | enrich(data) → {summaries, indexes, entities, links} | Code interpreter + RAG + NER | accuracy validation |
Phase 4 — Assembly & Packaging (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 4.1 Content Pyramid | plan_pyramid(pillar) → {15–25 assets, formats, deps} | LLM planner + dependency graph | plan approval |
| 4.2 Template Mapping | map_template(atoms, templates) → {filled_templates} | LLM + template engine | layout exceptions |
| 4.3 Modular Reassembly | reassemble(atoms, rules) → {configurations} | Constraint solver + LLM | novel config approval |
| 4.4 Sequencing | sequence(atoms) → {toc, order, rationale} | LLM + graph ranking | editorial override |
| 4.5 Cross-Referencing | link_atoms(atoms) → {internal_links, anchors} | Embeddings + LLM | link relevance check |
Phase 5 — Digital Download Generation (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 5.1 PDF Compilation | compile_pdf(content) → {pdf, cover, toc} | DOMPDF/Prince/Pandoc + AI cover gen | final proof |
| 5.2 EPUB Construction | build_epub(source) → {epub, spot_check} | Pandoc + epubcheck + LLM translation | stratified spot-check |
| 5.3 Gated Packaging | package_gated(product) → {optin_page, delivery_flow} | LLM copy gen + form API | marketing approval |
| 5.4 Marketplace Listing | list_product(product) → {copy, pricing, tags} | LLM + market data API | publish approval |
Phase 6 — Printable Product Generation (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 6.1 Data Prep | clean_data(csv) → {clean_data, schema, anomalies} | Code interpreter + LLM | schema change confirm |
| 6.2 Variable Template Design | design_template(schema) → {template+fields} | InDesign scripting / Canva API | design approval |
| 6.3 Merge & Preflight | merge_preflight(template, data) → {pdfs, report} | Data Merge + preflight tools | fix flagged issues |
| 6.4 Imposition & Batch | impose(pdfs) → {press_ready} | Imposition software + scripts | physical proof |
| 6.5 Print-Ready Export | export_print(imposed) → {pdfx, upload} | PDF/X validator + print API | final print approval |
Phase 7 — Post-Production & Distribution (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 7.1 QA | qa(outputs) → {defects, severity, fixes} | VLM (layout) + LLM (content) | critical defect |
| 7.2 Metadata Embedding | embed_metadata(product) → {title, rights, a11y_tags} | LLM + XMP/EPUB tools | rights verification |
| 7.3 Version Control | track_versions(changes) → {diffs, rollback} | Git/DVC + media API | major version approval |
| 7.4 Distribution Setup | distribute(product, channels) → {listings, uploads} | Marketplace APIs + RPA | compliance confirm |
| 7.5 Feedback Loop | optimize(perf_data) → {revision_tasks, ab_tests} | LLM + analytics | prioritization review |
Phase 8 — Estimation & Valuation (AI-Reframed)
| Block | AI Task | Model/Tool | HITL Trigger |
|---|---|---|---|
| 8.1 Cost Approach | estimate_cost(source) → {replacement_cost} | Regression + LLM | assumption validation |
| 8.2 Income Approach | forecast_income(product) → {dcf, projections} | ML forecasting | projection review |
| 8.3 Market Approach | find_comps(product) → {comps, valuation_range} | Embeddings + market DB | comp confirmation |
| 8.4 Max Size Estimation | estimate_size(content) → {uncompressed, compressed, page_count} | Code interpreter + compression tools | print-run approval |
Recommendation▾
The skill assumes familiarity with an external '37-block pipeline' that isn't defined here, making it incomplete as a standalone artifact; either inline the block definitions or clarify this is a companion/overlay skill with an explicit dependency.
End-to-End DAG
Scout→Rights→Quality→Aggregator
→ Atomizer→Qualifier→ContextReframer→Tagger→FormatInventory
→ WrapperGen→ChannelAdapter→FormatTranslator→BrandStyler→Enricher
→ PyramidPlanner→TemplateMapper→Assembler→Sequencer→Linker
→ [PDF/EPUB Compiler→Gated Packager→Marketplace Lister]
OR [DataCleaner→TemplateDesigner→Merge/Preflight→Imposition→PrintExporter]
→ QA→MetadataEmbedder→VersionController→DistributionAgent→FeedbackOptimizer
→ CostValuator→IncomeForecaster→MarketValuator→SizeEstimator
Best Practices
- Every block gets a confidence score — never let low-confidence output flow downstream silently.
- Keep human gates at: rights ambiguity, brand approval, high-cost print runs, final publish, financial assumptions.
- Use RAG/embeddings for anything requiring grounding (rights, comps, context repatriation) — pure generation hallucinates here.
- Version every atom independently; don't rebuild the whole product to fix one block.
Common Pitfalls
- Skipping the rights-audit gate to save time — creates legal exposure that automation can't undo.
- Letting the LLM self-score qualification (2.2) without an independent judge — inflates standalone-strength ratings.
- Automating print imposition (6.4) without a physical proof step — batch errors scale to the full print run.
- Treating valuation blocks (Phase 8) as deterministic — always output a range with confidence, not a single number.