An AI coding assistant skill to generate human-readable diagnostic reports from Airbyte connector sync logs.
Compatible with OpenAI Codex and Claude Code.
Parses Airbyte sync log files and produces structured reports with:
- Sync status, duration, and timing
- Source and destination versions
- Configuration summaries
- Per-stream record counts
- Destination write statistics by model
- State tracking (initial and final)
- Errors and warnings
Clone the skill into your project:
git clone https://github.com/faros-ai/analyze-airbyte-sync-logs-skill .agents/skills/analyze-sync-logsOr add as a submodule:
git submodule add https://github.com/faros-ai/analyze-airbyte-sync-logs-skill .agents/skills/analyze-sync-logsClaude Code looks for skills in .claude/skills/. Create a symlink:
mkdir -p .claude/skills
ln -s ../../.agents/skills/analyze-sync-logs .claude/skills/analyze-sync-logsAsk your AI assistant to analyze a sync log:
- "Analyze the sync log at
/path/to/sync_log.txt" - "What happened in this sync?"
- "Compare these two sync logs"
- "Were there any errors in this sync?"
The skill extracts structured data from the log file, which the AI assistant then uses to generate a human-readable report tailored to your question. Reports typically include:
- Sync status summary (success/failure, duration)
- Record counts per stream
- Destination write statistics
- Errors and warnings with context
- State changes between syncs
- Comparisons when analyzing multiple logs
| Question | What to Ask |
|---|---|
| Did the sync succeed? | "What's the sync status?" |
| How many records synced? | "Show records per stream" |
| Any errors? | "Were there any errors?" |
| Compare two syncs | "Compare these logs and highlight differences" |
| Check state preservation | "What's the initial and final state for X stream?" |
| Connector versions | "What source and destination versions were used?" |
Apache 2.0 - see LICENSE