session-searcher

1. Choose your approach:

   Path A — Direct examination (when session file is already known):

   • If a specific file path was provided, or context makes it obvious which session to examine, go directly to that file.
   • Read the session file and extract the relevant information.
   • Skip to step 3.

   Path B — Search first (when you need to discover sessions):

   • Run aichat search --json -n 10 "[query]" (use -g for global, --dir "project" to filter)
   • Parse results with jq to extract: session_id, project, created, snippet, file_path
   • Read session files at the file_path from results (max 3 files)
   • Run aichat search --help to see all options (date filters, branch filters, etc.)

2. Verify if needed: If referencing current codebase files (TO-DOS.md, etc.), read them to confirm current state

3. Summarize: Return a focused summary with key findings and session references

<output_format> Return a concise but comprehensive summary containing:

1. Key Findings: 2-3 bullet points answering the query
2. Relevant Sessions: Session IDs and dates for reference
3. Specific Content: Code snippets or quotes if directly relevant
4. Context: Brief explanation of how findings relate to query

Format as clean markdown, not raw JSON. </output_format>

Search: aichat search --json -n 10 "authentication design"

Summary:

Key Findings

• Session abc123 (Dec 10): Discussed JWT vs session-based auth, decided on JWT for API
• Session def456 (Dec 8): Implemented refresh token rotation pattern