phoenixclaw_skill

Overview

This skill passively scans daily conversations and session logs on a schedule or on demand to produce polished Markdown journals. It uses semantic understanding to identify journal-worthy moments, extract images, and surface patterns and growth opportunities. Journals include plugin-generated sections, image embeddings, timeline updates, and profile evolution notes. The pipeline is deterministic and must run end-to-end to guarantee complete results.

How this skill works

On each run the skill loads user configuration, reads memory files and all session logs for the target window, extracts text and media, then identifies moments (decisions, emotions, milestones) and matches images to those moments. It runs registered plugins at defined hook points (pre-analysis, post-moment-analysis, post-pattern-analysis, journal-generation, post-journal) so Ledger, analytics, and custom sections can contribute. Finally it synthesizes a Markdown journal from templates, updates timeline/growth/profile files, and copies curated assets into the entry folder.

When to use it

• Ask: "Show me my journal" or "Generate today's journal" for on-demand reconstruction.
• Request pattern analysis: "Analyze my patterns" or "How am I doing?" to surface trends and mood changes.
• Create summaries: "Generate weekly summary" or "Monthly summary" to compile high-level insights.
• Regenerate entries after adding memories or correcting assets: "Regenerate my journal".
• Rely on nightly cron to passively produce the day's journal without manual input.

Best practices

• Always run the full 9-step workflow; partial runs omit images, ledger entries, or plugin output.
• Keep memory/YYYY-MM-DD.md files for explicit reflections—these are combined with session logs.
• Enable rolling window scheduling to avoid midnight spillover losses for late-night activity.
• Filter out heartbeat and cron noise when scanning session logs to avoid false moments.
• Mark finance screenshots for Ledger routing and exclude them from the main narrative unless explicitly relevant.

Example use cases

• Daily automatic journal generation at 22:00 local time with embedded highlight photos.
• On-demand regeneration after adding a missing memory file or newly copied images.
• Weekly mood and pattern report that updates growth-map.md and suggests skill recommendations.
• Ledger-enabled flow that routes payment screenshots to financial records while keeping journals clean.
• Timeline integration that appends notable events to a master timeline for long-term review.