Org Context
Organizational taxonomy gives every command access to your org's vocabulary, enabling validation and shorthand resolution.
The Goal
Transform informal references into validated entities:
User: "write a story for the billing module in WebApp"
↓ Claude resolves
Product: WebApp (Core SaaS platform)
Module: Billing (Payment processing and invoicing)
Without org context, Claude guesses. With it, Claude validates against your actual taxonomy and uses canonical names.
Taxonomy Categories
Six taxonomy types managed via forge-lib:
| Category | Purpose | Example Values |
|---|---|---|
| Products | Product lines | WebApp, MobileApp, API Platform |
| Modules | Functional areas | Authentication, Billing, Notifications |
| Systems | Infrastructure | API Gateway, Data Warehouse |
| Clients | Key customers | Acme Corp, Globex Industries |
| Teams | Organizational units | Platform Team, Finance Team |
| Integrations | External systems | Salesforce, Stripe, Twilio |
Taxonomy Query Pattern
Commands query taxonomy via forge-lib:
forge memory get-taxonomy products --directory . forge memory get-taxonomy modules --directory . forge memory get-taxonomy clients --directory . forge memory get-taxonomy teams --directory . forge memory get-taxonomy integrations --directory . forge memory get-taxonomy systems --directory .
Returns JSON arrays of valid values for validation and suggestion.
Resolution Strategy
1. Shorthand Resolution
Map informal references to canonical names:
- •"the mobile app" → MobileApp
- •"Acme" → Acme Corp
- •"billing stuff" → Billing (module)
Query taxonomy via forge-lib to validate and resolve.
2. Validation
When a command field requires a taxonomy value:
- •Query forge-lib for valid values
- •Match user input (fuzzy, case-insensitive)
- •If ambiguous, ask user to clarify
- •If unknown, offer to add via
/memory:setup-org
3. Suggestion
When gathering input for a field:
What product is this for? Your products: - WebApp (Core SaaS platform) - MobileApp (Field operations app) - API Platform (Developer tools)
Query taxonomy for current values and display as suggestions.
Missing Taxonomy Handling
If taxonomy doesn't exist:
- •Accept freeform values (don't block the user)
- •Inform: "Run
/memory:setup-orgto configure your taxonomy" - •After workflow: "I used '[value]'. Want to add it to your taxonomy?"
If value not in taxonomy:
- •Flag: "I don't see '[value]' in your [category] taxonomy"
- •Offer: "Should I add it? (Yes / Use it anyway / Enter different value)"
- •On confirmation: Call
forge memory set-taxonomy {taxonomy_type} --add "{value}" --directory .
Integration with Commands
Commands that reference org entities should:
1. Query Taxonomy:
forge memory get-taxonomy products --directory .
2. Present Options:
Which product? 1. WebApp 2. MobileApp 3. API Platform [Enter number or name]
3. Validate Input: Match against returned taxonomy (case-insensitive, fuzzy).
4. Handle Unknown: Offer to add new value or accept freeform.
Cross-Plugin Usage
Product Forge, Tasks Forge, and Report Forge all query the same taxonomy:
- •Product Forge: Product, module, client fields on cards
- •Tasks Forge: Related product/module for tasks
- •Report Forge: Scope reports to specific products/modules
Shared taxonomy ensures consistency across plugins.
Bootstrapping
Use /memory:setup-org to populate taxonomy through interactive interview:
- •Products and modules
- •Systems
- •Clients
- •Integrations
- •Teams
Re-runnable—loads existing values and asks what to update.
Update Flow
Adding a value:
forge memory set-taxonomy products --add "New Product" --directory .
Removing a value:
forge memory set-taxonomy products --remove "Old Product" --directory .
Reasoning Principles
1. Fuzzy Matching
Accept "mobile", "mobile app", "the mobile app" → all resolve to "MobileApp"
2. Growth Over Validation
Don't reject unknown values—offer to add them. Taxonomy grows organically.
3. Context Enrichment
Use taxonomy to enrich card creation:
- •User says "billing"
- •Taxonomy provides full description: "Billing (Payment processing and invoicing)"
- •Card frontmatter gets enriched value
4. Progressive Setup
Don't require taxonomy up-front. Accept freeform values, suggest taxonomy setup after seeing patterns.
All taxonomy operations delegated to forge-lib—this skill provides the resolution strategy only.