Industry Catalog

Context Independence

This skill reads ALL required state from pursuit project files and catalog directories — it does not depend on prior conversation context. If invoked after trends-resume or value-modeler, context compaction is safe and recommended before starting.

Before executing any operation, run /compact to free working memory. Catalog import operations read the full tips-value-model.json and merge it with existing catalog entities, which benefits from maximum available context.

If /compact is unavailable or this is the first skill in the session, skip compaction and proceed directly.

Manage persistent industry catalogs — the reusable knowledge base that makes each successive customer pursuit faster, more consistent, and more valuable.

Why Catalogs Matter

Without catalogs, every TIPS pursuit starts from scratch. The trend-scout generates candidates, the value-modeler builds paths and solutions, but all that structured knowledge lives only in the pursuit directory. The next customer engagement in the same industry repeats the same work.

Catalogs solve this by creating a persistent, curated repository per industry/subsector. When a new pursuit is initialized, it inherits relevant TIPs, Solution Templates, SPIs, and Metrics from the catalog — giving the consultant a head start and ensuring consistency across engagements.

This implements the centralized repository concept from the TIPS Value Modeler methodology (WO2018046399A1, Claim 3 and Fig. 4).

Catalog Structure

cogni-trends/catalogs/{industry}/{subsector}/
├── catalog.json              # Root manifest (industry, version, stats)
├── tips-entities.json        # Curated TIP entities (T, I, P roles)
├── solution-templates.json   # Proven Solution Templates
├── spis.json                 # Validated Solution Process Improvements
├── metrics.json              # Effective success KPIs
├── collaterals.json          # Available supporting content
└── .history/                 # Version snapshots (catalog.json copies)

Catalogs live at the plugin workspace level (cogni-trends/catalogs/), not inside any pursuit project directory. They persist across projects and sessions.

Operations

init — Create a New Catalog

Create an empty catalog for an industry/subsector combination.

/trends-catalog init

Steps:

1. Ask user for industry and subsector (use the same taxonomy as trend-scout: read $CLAUDE_PLUGIN_ROOT/skills/trend-scout/references/industry-taxonomy.md)
2. Optionally load a taxonomy template as seed data (see Taxonomy Templates below)
3. Create the directory structure and catalog.json

catalog.json schema:

{
  "version": "1.0.0",
  "industry": {
    "primary": "manufacturing",
    "primary_en": "Manufacturing",
    "primary_de": "Fertigung",
    "subsector": "automotive",
    "subsector_en": "Automotive",
    "subsector_de": "Automobil"
  },
  "created": "2026-03-11T10:00:00Z",
  "updated": "2026-03-11T10:00:00Z",
  "stats": {
    "tips_entities": 0,
    "solution_templates": 0,
    "spis": 0,
    "metrics": 0,
    "collaterals": 0,
    "pursuits_contributed": 0
  },
  "taxonomy_template": null,
  "pursuit_history": []
}

import — Promote Pursuit Data to Catalog

Import curated entities from a completed value-modeler pursuit into the catalog. This is the write-back loop that makes catalogs grow smarter over time.

/trends-catalog import

Steps:

1. Discover TIPS projects with completed value models (workflow_state: "curated" or "complete")
2. If curation recommendations exist (tips-value-model.json → curation_recommendations), use them as the starting point — these are pre-analyzed by Phase 5
3. If no curation recommendations exist, analyze the value model directly (same criteria as Phase 5: paths with BR >= 4.0, Tier 1-2 STs, etc.)
4. Find the matching catalog (by industry/subsector) or offer to create one
5. Present import candidates to user for approval — show what will be added/merged
6. On approval, write entities to catalog files

Import rules:

• Generalize before importing. Strip customer names, specific product references, and pursuit-specific context. The catalog should contain industry-general patterns.
• Deduplicate. Check if a semantically similar entity already exists in the catalog. If so, offer to merge (strengthen the existing entry) rather than create a duplicate.
• Track provenance. Each imported entity gets a provenance field:

{
  "source_pursuit": "automotive-ai-predictive-maintenance-abc12345",
  "imported_at": "2026-03-11T14:00:00Z",
  "original_ref": "st-001",
  "generalization_notes": "Replaced 'BMW Regensburg line sensors' with 'production line sensors'"
}

• Update stats. Increment stats counts and add pursuit to pursuit_history:

{
  "pursuit_slug": "automotive-ai-predictive-maintenance-abc12345",
  "imported_at": "2026-03-11T14:00:00Z",
  "entities_imported": { "tips": 3, "sts": 2, "spis": 1, "metrics": 2, "collaterals": 0 }
}

list — Show Available Catalogs

/trends-catalog list

List all catalogs in cogni-trends/catalogs/ with their industry, subsector, entity counts, and last update date. Show a compact table.

show — Display Catalog Contents

/trends-catalog show [industry/subsector]

Display the full contents of a specific catalog with entity counts per type, most recent imports, and coverage summary. If the catalog was initialized from a taxonomy template, show coverage against the template categories.

analytics — Cross-Pursuit Insights

/trends-catalog analytics [industry/subsector]

Analyze patterns across all pursuits that contributed to a catalog:

• Trend frequency: Which TIP entities appear in multiple pursuits? These are market signals.
• Solution popularity: Which STs are most frequently selected? These are portfolio priorities.
• BR distribution: How do BR scores distribute across pursuits for the same entities?
• Coverage gaps: Which catalog areas have few or no entities?
• Maturity curve: How has the catalog grown over time?

Present insights as a structured summary with actionable recommendations.

Entity Schemas

Catalog TIP Entity

TIP entities in the catalog are generalized versions of pursuit-specific candidates. They retain the TIPS role (T/I/P) and dimension mapping but are stripped of pursuit context.

{
  "entity_id": "cat-tip-001",
  "name": "Increasing regulatory pressure on AI systems",
  "description": "Governments worldwide are implementing AI governance frameworks that create compliance requirements for industrial AI deployments",
  "tips_role": "T",
  "dimension": "externe-effekte",
  "subcategory": "regulierung",
  "keywords": ["ai-regulation", "compliance", "governance"],
  "horizon": "act",
  "provenance": {
    "source_pursuit": "automotive-ai-predictive-maintenance-abc12345",
    "imported_at": "2026-03-11T14:00:00Z",
    "original_ref": "externe-effekte/act/1",
    "generalization_notes": "Broadened from EU AI Act to global AI governance"
  },
  "pursuit_appearances": 3,
  "avg_business_relevance": 4.2
}

Catalog Solution Template

{
  "st_id": "cat-st-001",
  "name": "Predictive Quality Analytics Platform",
  "description": "ML-based quality prediction integrated with production line sensor data for real-time defect detection",
  "category": "software",
  "enabler_type": "process_improvement",
  "typical_linked_tips": ["cat-tip-001", "cat-tip-005", "cat-tip-012"],
  "generation_mode": "portfolio-anchored",
  "blueprint_summary": {
    "block_count": 3,
    "taxonomy_span": [1, 6, 7],
    "taxonomy_depth": 3,
    "source_readiness_score": 0.64
  },
  "portfolio_mapping_hint": {
    "suggested_dimension": 6,
    "suggested_category": "6.6",
    "suggested_category_name": "AI, Data & Analytics"
  },
  "provenance": {
    "source_pursuit": "automotive-ai-predictive-maintenance-abc12345",
    "imported_at": "2026-03-11T14:00:00Z",
    "original_ref": "st-001",
    "generalization_notes": "Removed brand-specific sensor references"
  },
  "pursuit_appearances": 2,
  "avg_ranking_value": 4.1
}

The portfolio_mapping_hint uses b2b-ict-portfolio taxonomy category IDs when a taxonomy template is loaded. This enables downstream mapping to cogni-portfolio features.

Blueprint Field Preservation

When importing Solution Templates that have solution_blueprint and generation_mode fields (introduced in the multi-dimensional solutioning model), preserve this data in the catalog entry. The blueprint captures the full portfolio composition needed to deliver the solution — building blocks mapped to taxonomy categories with role-weighted readiness scoring.

• generation_mode: Carry forward as-is ("portfolio-anchored" or "abstract")
• solution_blueprint: Generalize building block references (strip pursuit-specific feature slugs) but preserve the dimensional structure (roles, taxonomy categories, coverage levels)
• readiness_score: Store as a reference point but note it reflects the source pursuit's portfolio state — readiness will differ for a new pursuit with a different portfolio

Solution Templates without blueprints (older projects) continue to work — blueprint fields are optional.

Catalog SPI

{
  "spi_id": "cat-spi-001",
  "name": "Establish data governance policy for sensor data",
  "description": "Define data ownership, quality standards, and access controls for production data",
  "change_type": "governance",
  "typical_st_refs": ["cat-st-001"],
  "provenance": { "..." : "..." },
  "pursuit_appearances": 2
}

Catalog Metric

{
  "metric_id": "cat-met-001",
  "name": "Defect rate reduction",
  "unit": "percentage",
  "direction": "decrease",
  "typical_baseline_range": "2-8%",
  "typical_target_range": "0.5-2%",
  "provenance": { "..." : "..." },
  "pursuit_appearances": 3
}

Catalog Collateral

{
  "collateral_id": "cat-col-001",
  "name": "Predictive Maintenance ROI Case Study",
  "type": "case-study",
  "typical_st_refs": ["cat-st-001"],
  "status": "exists",
  "provenance": { "..." : "..." }
}

Taxonomy Templates

Catalogs can be initialized from a taxonomy template that pre-defines the category structure. The template does not provide entities — it provides the dimensional framework that guides what kinds of entities should be collected.

Currently supported template:

Template	Source	Categories
b2b-ict-portfolio	cogni-trends	57 categories across 8 dimensions (0-7)

When a taxonomy is loaded, the catalog gains:

• A taxonomy_template field in catalog.json pointing to the template
• Coverage tracking: which taxonomy categories have catalog entities mapped to them
• Gap analysis: which categories remain empty

To use: during /trends-catalog init, when asked about taxonomy, reference the b2b-ict-portfolio framework. Read the taxonomy definition from: $CLAUDE_PLUGIN_ROOT/references/taxonomies/b2b-ict-portfolio.md

If the file does not exist yet, the user can provide the taxonomy source or skip taxonomy loading.

Integration with Value Modeler

Phase 0 Enhancement: Load Catalog

When value-modeler Phase 0 discovers a project, it should also search for a matching industry catalog:

1. Read the project's industry/subsector from tips-project.json
2. Check cogni-trends/catalogs/{industry}/{subsector}/catalog.json
3. If found, report: "Found {industry}/{subsector} catalog with X entities. Catalog data will inform relationship building and solution generation."
4. Store catalog path in .metadata/value-modeler-output.json as catalog_path

The catalog does NOT replace trend-scout candidates. It supplements them:

• Phase 1 can use catalog TIP entities as seed patterns for path building
• Phase 2 can pre-populate Solution Templates from catalog STs
• Phase 2 can suggest SPIs and Metrics from the catalog

Phase 5 Enhancement: Write-Back

When Phase 5 curation completes and generates recommendations:

• If a matching catalog exists, offer to run /trends-catalog import immediately
• If no catalog exists, suggest creating one with /trends-catalog init
• The curation recommendations become the import candidate list

Language Support

This skill follows the shared language resolution pattern — see $CLAUDE_PLUGIN_ROOT/references/language-resolution.md.

Interaction language: Read workspace language from .workspace-config.json (via ${PROJECT_AGENTS_OPS_ROOT}/.workspace-config.json or CWD) at startup. Use this for all user-facing messages — catalog operation prompts, status displays, import approval questions, and analytics summaries.

Output language: Catalog entity names and descriptions can be bilingual (matching the pursuit language). The industry object always carries both _en and _de variants for consistency with the industry taxonomy.

Output

All catalog operations report results to the user with clear counts and summaries. Import operations always require explicit user approval before writing.