Quick Reference

• File: SKILL.md (295 lines)
• Type: Main entry point and router
• Input: 10-parameter config object
• Output: Routes to appropriate skills

The orchestrator accepts a config object with 10 parameters:

#	Parameter	Type	Required	Default
1	DOC_TYPE	enum	Yes	all
2	FORMAT	enum	Yes	astro
3	SCOPE	enum	Yes	full
4	FOCUS_TARGET	string	No	null
5	LANGUAGE	string	Yes	auto-detect
6	I18N	boolean	No	false
7	RECORD	boolean	No	false
8	PROJECT_PATH	path	Yes	workspace
9	SEO	boolean	No	true
10	LLM_OPTIMIZE	boolean	No	true

graph TB

    D{"DOC_TYPE?"}
    K["knowledge\npersona + jtbd + flow"]
    T["tech\narchitecture + DB + deploy"]
    S["sop\nknowledge first, then SOPs"]
    A["api\nendpoint reference"]
    ALL["all\nknowledge → tech → sop → api"]

    D -->|"knowledge"| K
    D -->|"tech"| T
    D -->|"sop"| S
    D -->|"api"| A
    D -->|"all"| ALL

Routing summary: The orchestrator maps DOC_TYPE to the corresponding skill files. For all, it runs all types sequentially: knowledge → tech → sop → api. For sop, it auto-runs knowledge first if not already generated.

Step	Action	Condition
1	Gather input	Always
1b	Generate execution plan	Always
2	Analyze codebase	Always
3	Apply content guidelines	Always
3b	Apply SEO + LLM guidelines	If SEO=yes or LLM=yes
4	Generate documents	Based on DOC_TYPE
5	Export	Based on FORMAT
5b	Generate sitemap	If SEO=yes
5c	Run SEO audit	If SEO=yes
6	Summary	Always

• analyze-codebase API
• Skill pipeline workflow
• System Architecture