From tasktrace-mcp
This skill should be used when the user asks about what they know, what they have read, what they have learned, or what is in their notes, documents, or knowledge base. Trigger phrases include "what do I know about X", "find notes on X", "what have I read about X", "what's in my knowledge base", "find claims about X", "what communities relate to X", "what does the graph know", "summarize what I know", and any question where the answer lives in the user's curated knowledge rather than in their activity history. Powered by TaskTrace's graph RAG over the user's currently selected knowledge directory.
How this skill is triggered — by the user, by Claude, or both
Slash command
/tasktrace-mcp:tasktrace-knowledgeThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
TaskTrace builds a knowledge graph — communities, nodes, edges, and claims — over a directory of the user's content that they selected inside TaskTrace. The graph mixes everything TaskTrace has ingested for that directory: notes the user wrote, documents they read, and anything extracted from screen captures and transcripts that got promoted into persistent knowledge. Treat the graph as a singl...
TaskTrace builds a knowledge graph — communities, nodes, edges, and claims — over a directory of the user's content that they selected inside TaskTrace. The graph mixes everything TaskTrace has ingested for that directory: notes the user wrote, documents they read, and anything extracted from screen captures and transcripts that got promoted into persistent knowledge. Treat the graph as a single source: it answers "what does the user know or believe", not "what did the user do".
The tasktrace_graph_search tool queries this graph. It is separate from tasktrace_search, which ranks activity history.
Use tasktrace_graph_search when the question is about knowledge:
Use tasktrace_search (see the tasktrace-context skill) when the question is about activity:
If the user's phrasing is ambiguous — "what do I have on X" — run both. The tools are cheap and cover different surfaces. Present the combined picture.
Do not try to distinguish for the user whether a piece of knowledge came from their notes, a PDF they read, or text extracted from a screenshot. The graph treats these uniformly on purpose. Answer from the graph content without apologizing for the provenance.
tasktrace_graph_search input:
{
"query": "string (required)",
"limit": "number (optional, default 3, clamped to [1, 10])"
}
query is natural language. Pass the user's phrasing through, cleaned of filler.limit is the number of reranked graph hits to keep. Default 3 is narrow by design; raise to 5–10 for exploratory questions ("everything I know about X").GraphRAGRetrievalResult. Surface the content directly; the structure contains enough context (community summaries, node text, claim statements) that summarization can work from it without external lookup.If tasktrace_graph_search returns no hits, the user may not have selected a knowledge directory in TaskTrace, or the selected directory has not been indexed. State that briefly — "the knowledge graph didn't return hits for that" — and offer to fall back to activity search if the question plausibly could be answered from recent work.
tasktrace_search over activity./tasktrace-mcp:setup if this persists.query: the tool rejects it. Validate input before calling.npx claudepluginhub warrenronsiek/tasktracemcpplugin --plugin tasktrace-mcpTraverses knowledge graph across memories, entities, and relationships for comprehensive context. Use before planning complex work, investigating concept connections, or answering 'what do you know about X'.
Searches an Obsidian vault for prior knowledge, past decisions, and accumulated learnings, surfacing relevant notes before answering from general knowledge. Uses index-first retrieval with fallback content search.
Queries and manages a project knowledge graph across tasks, SOPs, memories, and concepts. Captures patterns, pitfalls, decisions, and learnings for persistent retrieval.