Evidence backed semantic storage demos for AI agents, with reproducible scenarios and inspectable accepted evidence.
-
Updated
Sep 8, 2026 - Python
Evidence backed semantic storage demos for AI agents, with reproducible scenarios and inspectable accepted evidence.
Independent Claude Code plugin inspired by Stanford STORM: contradiction-aware multi-agent evidence review with source mapping, disagreement tracking, and decision-ready briefs.
面向 OpenClaw / Codex Agent 的决策导向 deep research skill,强调任务路由、证据可追溯、current-state verification、反证约束与可审计交付。
Full-stack, self-hosted memory infrastructure for AI agents: algorithms, traceable recall, knowledge graph, API/SDK/MCP, apps, deployment, benchmarks, and custom-memory adapters.
Paste-only Streamlit research briefs with ordered evidence, deterministic exports, and optional gated AI summarization.
Codebase and evaluation framework for my Bachelor's Thesis (TCC), focused on developing a Relational Retrieval-Augmented Generation (RAG) system that combines vector search, cross-document entity linking, and explicit source citation to minimize LLM hallucinations
Turn screenshot batches into sourced action items, dates, duplicate groups, and review-ready calendar drafts.
To associate your repository with the source-traceability topic, visit your repo's landing page and select "manage topics."