head-tea/ChatMemory
Skill Claude CodeCodex
Export WeChat/QQ chat records via WeFlow (WeChat) or QCE+NapCatQQ (QQ) HTTP APIs. Full pipeline: export → clean → knowledge graph (Obsidian vault) → NotebookLM analysis. Sub-commands: /chatmemory-wechat, /chatmemory-qq, /chatmemory-obsidian, /chatmemory-obsidian-then-notebooklm, /chatmemory-obsidian-upgrade (v1.3).…
head-tea/ChatMemory
Skill Claude CodeCodex
Full automated pipeline: clean chat records → build Obsidian knowledge graph vault → auto-upload to NotebookLM with KG enrichment. Combines --obsidian and --notebooklm-enrich in one command. Use when the user wants the complete end-to-end KG + NotebookLM workflow.
head-tea/ChatMemory
Skill Claude CodeCodex
Update an existing project/chat to the Obsidian knowledge graph vault. Runs the full KG pipeline (candidate extraction → Claude analysis → calibration → vault generation) without producing NotebookLM output. Equivalent to chatcleaner.py --obsidian-only. Use when you want to refresh a project's knowledge graph without…
head-tea/ChatMemory
Skill Claude CodeCodex
Generate an Obsidian knowledge graph vault from cleaned chat records. Runs candidate extraction, Claude-powered semantic analysis, calibration, contradiction detection, and vault file generation (v1.3). Turns chat discussions into a navigable knowledge network with 17 colored relationship types and dual-axis…
head-tea/ChatMemory
Skill Claude CodeCodex
Export QQ chat records via QCE + NapCatQQ (OneBot HTTP API). Auto-launches QCE, exports chats, then hands off to the shared ChatMemory pipeline (cleaning → knowledge graph → NotebookLM). Use /chatmemory-qq for QQ-specific export workflows.
head-tea/ChatMemory
Skill Claude CodeCodex
Export WeChat chat records via WeFlow HTTP API. Auto-launches WeFlow, exports chats, then hands off to the shared ChatMemory pipeline (cleaning → knowledge graph → NotebookLM). Use /chatmemory-wechat for WeChat-specific export workflows.
head-tea/ChatMemory
Skill Claude CodeCodex
Find domain experts from the ChatMemory Obsidian knowledge graph. Searches person-type entities ranked by referencecount and relationship connections. Use when the user asks "who knows X", "find expert for Y", or "who should I ask about Z".
head-tea/ChatMemory
Skill Claude CodeCodex
Assemble a structured project background document from the ChatMemory Obsidian knowledge graph. Extracts relevant methodologies, historical decisions, known pitfalls, domain experts, and knowledge gaps for a given project description. Use when starting a new project or evaluating a technical approach.
head-tea/ChatMemory
Skill Claude CodeCodex
Search and retrieve knowledge from the ChatMemory Obsidian knowledge graph vault. Supports typed traversal, hybrid recall, expert finding, gap detection, and project context assembly. Use when the user asks to search the knowledge graph, find experts, check knowledge gaps, or assemble project background.
head-tea/ChatMemory
Skill Claude CodeCodex
Detect knowledge gaps and weaknesses in the ChatMemory Obsidian knowledge graph. Scans for low-confidence, low-coverage entities and topics with insufficient discussion depth. Use when the user asks "what are we missing", "knowledge gaps", or "what needs more research".
head-tea/ChatMemory
Skill Claude CodeCodex
Two-tier review of uncertain entity merges between WeChat and QQ knowledge graphs. Tier 1: Claude AI auto-reviews candidates using semantic understanding — high-confidence decisions applied automatically. Tier 2: Only ambiguous cases are escalated to the user for manual approval. Use when vaultmerger flags uncertain…
head-tea/ChatMemory
Skill Claude CodeCodex
A seven-phase workflow for developing a new ChatMemory version, from setup and discussion through research, implementation, audit, and release synchronization.