chat-history-searchlisted

Exhaustive search across ALL of Bryce's local Claude chat history, both Cowork and the Claude Code CLI, to find past prompts, recover a conversation, inventory how often a pattern was used, or audit prior work across projects. Knows every log location, every JSONL line shape, and the false-positive gotchas that trip up naive grep.
BryceEWatson/claude-global-skills · ★ 0 · AI & Automation · score 68

Install: claude install-skill BryceEWatson/claude-global-skills

# Chat History Search Exhaustive search across the two separate corpora of local Claude chat history. ## Why this skill exists The first time I tried this, I missed the entire Cowork corpus (1,674 JSONL files at `%AppData%\Claude\local-agent-mode-sessions\`) because I only searched `~/.claude/projects/`. The user had to point me at their blog post documenting both locations. I also produced false positives by reporting TodoWrite items, `<task-notification>` wrappers, and document content read via tool_results as if they were Bryce's typed prompts. This skill encodes the full corpus map plus the verification rules that catch those false positives — so the next search is right the first time. ## When this skill triggers User intent that maps here: - "find all uses of X in my chat history" / "every time I asked Claude to Y" - "when did I last Z" / "inventory my prompts about W" - "recover that conversation where we talked about V" - "audit my sessions for U" - "search my local logs for T" ## Storage map — search ALL of these ### Corpus 1: Claude Code CLI / VS Code (`~/.claude/projects/`) ``` %USERPROFILE%\.claude\projects\<encoded-cwd>\<sessionId>.jsonl %USERPROFILE%\.claude\projects\<encoded-cwd>\<sessionId>\subagents\agent-*.jsonl ``` `<encoded-cwd>` is the working dir path with `:` and `/` replaced by `-`, lowercased (e.g. `c--Users-Bryce-Projects-brycewatson-com`). One project = one directory; one session = one JSONL file. Subagent transcripts live in a sibling dire