codex-history-ingest

To convert Codex CLI session history into an Obsidian wiki, this Skill parses local JSONL logs, redacts private data, clusters topics, and generates searchable wiki pages. It distills noisy conversation rollouts into durable, structured knowledge base pages.

Yes, you can incrementally sync Codex conversation logs without duplicating wiki pages by using append mode. This mode utilizes manifest-based delta detection to process only new session data, updating existing topic clusters and vault index files without a full rebuild.

Topic clustering when ingesting coding session logs groups extracted knowledge into structured categories like projects, concepts, and skills. Instead of creating one page per session, it synthesizes information across multiple sessions to build cohesive wiki pages.

Yes, your private data is redacted when exporting Codex session transcripts to a wiki. The ingestion process applies privacy-first distillation by actively removing secrets and summarizing sensitive content rather than copying raw conversation transcripts directly.

The best way to rebuild an Obsidian knowledge base from past coding sessions is a full synchronization workflow. This mode re-ingests all available local history, completely re-clustering topics and overwriting existing wiki pages to ensure a clean, comprehensive knowledge base.

Yes, Codex history ingestion skips low-signal telemetry when parsing JSONL logs. It selectively mines session rollouts and index data, filtering out noisy telemetry artifacts to ensure only durable, reusable knowledge is written to your Obsidian wiki.