wiki-agent

Extract topic-specific information from an AI agent's conversation history into wiki pages.

4|Updated May 9, 2026

Search Tags:#obsidian #synthesis #wiki #cross-agent #history search #knowledge ingestion #query-driven

Ready to run

npx skills add https://github.com/itsmarsss/memento --skill wiki-agent

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: wiki-agent
Source: https://github.com/itsmarsss/memento/tree/main/obsidian-wiki/.cursor/skills/wiki-agent
Command: npx skills add https://github.com/itsmarsss/memento --skill wiki-agent

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Users need fast, topic-specific knowledge from a different AI agent’s past work without bulk-ingesting everything they did.

Core Features & Use Cases

Targeted cross-agent search: Finds relevant sessions in a chosen agent’s raw history for a specific topic or query.
Selective extraction and distillation: Ingests only the relevant blobs, then updates or creates the appropriate wiki pages.
Immediate synthesized answer: Returns an actionable summary usable in the user’s current session, including what was ingested and where it landed in the wiki.
Flexible invocation modes: Uses query mode for precise retrieval or recent unprocessed sessions mode when no query is provided.

Quick Start

Run /wiki-claude "how did I set up auth middleware" to search Claude’s history for that topic, ingest the relevant parts into the wiki, and get a synthesized answer immediately.

Frequently Asked Questions about wiki-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I pull specific knowledge from another AI agent's conversation history?▼

Extract targeted knowledge from another agent's history by querying a specific topic, scoring its past sessions, and ingesting only the relevant portions into your wiki. This provides an immediate synthesized answer for cross-agent debugging and knowledge transfer without bulk-importing raw history.

What's the best way to transfer how a different tool previously solved a problem into my current session?▼

Transfer previous solutions by performing selective extraction and distillation from the other agent's raw history. This ingests only relevant blobs, updates your wiki pages with tracked sources, and returns an actionable synthesized summary directly into your current working session.

Can I search another agent's raw history without importing all of its past conversations?▼

Yes, you can search another agent's raw history without bulk importing by using a query-driven approach. It efficiently reads the agent history inventory, scores and selects the top relevant sessions, and performs targeted extraction to merge only necessary information into your wiki.

How does cross-agent knowledge ingestion work when updating wiki pages?▼

Cross-agent knowledge ingestion works by retrieving topic-specific information from a chosen agent, extracting the relevant portions, and merging or updating existing wiki pages with tracked sources. This ensures your wiki remains current with cross-agent insights while maintaining source provenance.

What happens if I don't have a specific query for cross-agent history search?▼

If you lack a specific query for cross-agent history search, the system uses a recent unprocessed sessions mode. This flexible invocation mode automatically evaluates newer history entries that have not yet been ingested, ensuring ongoing knowledge transfer even without a targeted prompt.