codex-history-ingest
Ingest Codex CLI conversation history into the Obsidian wiki. Use this skill when the user wants to mine their past Codex sessions for knowledge, import their ~/.codex folder, extract insights from previous coding sessions, or says things like "process my Codex history", "add my Codex conversations to the wiki", or "what have I discussed in Codex before". Also triggers when the user mentions .codex sessions, rollout files, session_index.jsonl, or Codex transcript logs.
Security Assessment
Detected risks:
About codex-history-ingest
The codex-history-ingest skill mines a user's past Codex CLI sessions and distills durable knowledge into an Obsidian wiki. It targets the fact that session logs are rich but noisy, focusing extraction on reusable knowledge rather than operational telemetry.
It resolves configuration (the Obsidian vault path and a Codex history path that defaults to ~/.codex), supports append mode (processing only new or modified files tracked in a manifest) and full re-ingest, and parses Codex data, namely the session_index.jsonl inventory and sessions/**/rollout-*.jsonl event logs, with rules that prioritize user intent and assistant outputs while filtering token-accounting and tool plumbing. It applies a critical privacy filter (removing secrets and not quoting raw system or developer prompts), clusters knowledge by stable topic rather than per session, and writes distilled wiki pages with summary frontmatter, confidence and lifecycle fields, and provenance markers, then updates the manifest, index, and log and optionally refreshes a QMD search index.
It is most useful for developers who want to turn accumulated Codex conversations into an organized, deduplicated knowledge base in Obsidian.
FAQ
What does this skill do?
It ingests Codex CLI conversation history from ~/.codex and distills durable knowledge into organized Obsidian wiki pages, rather than dumping raw transcripts.
How does it avoid re-processing everything?
Append mode uses a .manifest.json to process only new files or those modified since the last ingest; full mode reprocesses everything when explicitly requested.
Which Codex data does it read?
Primarily session_index.jsonl for the inventory and sessions/**/rollout-*.jsonl for structured transcript events, with history.jsonl as an optional fallback; it avoids SQLite internals.
How does it handle privacy?
It removes API keys, tokens, passwords, and credentials, redacts private identifiers, and summarizes rather than quoting raw system/developer prompts or transcripts.
How is the knowledge organized?
Clustered by stable topics rather than per session, and routed into projects, concepts, skills, entities, and synthesis pages with summary, confidence/lifecycle, and provenance metadata.
Install codex-history-ingest
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
ar9av/obsidian-wiki