Inspect and diagnose AutoRun pipeline failures via snapshots and audit logs.
npx skills add https://github.com/mcleanT/AutoReview --skill debug-pipeline-mcleant
Please help me install this Agent Skill. Skill: debug-pipeline Source: https://github.com/mcleanT/AutoReview/tree/main/.claude/skills/debug-pipeline Command: npx skills add https://github.com/mcleanT/AutoReview --skill debug-pipeline-mcleant
This skill provides step-by-step guidance and commands to inspect, diagnose, and recover AutoReview pipeline failures by loading snapshots, examining node state and audit logs, identifying common errors, and resuming execution from the last good state.
Load the latest snapshot at output/snapshots/latest.json, inspect phase, candidate and extraction counts, check the last audit log entries, apply fixes, and resume from the failing or chosen node.
High-intent search queries and answers about installing and using this skill.
To troubleshoot an AutoReview pipeline failure, load the latest KnowledgeBase snapshot at output/snapshots/latest.json, inspect the current_phase and candidate counts, and review node audit logs to identify errors to pinpoint the failing DAG node.
The best way to resume an LLM pipeline run after a TokenBudgetExceeded error is to load the last good snapshot, adjust the token budget configuration, and restart execution from the failing node via the CLI or Python API.
Yes, you can diagnose DAG node failures by reviewing audit logs and analyzing token consumption across all 15 nodes to map common failure modes like API errors to their specific fixes.
Yes, you need configured LLM and service API keys to resume pipeline execution and re-run downstream nodes, alongside access to output/snapshot files and the KnowledgeBase load_snapshot/save_snapshot APIs.
Your pipeline may fail during section_writing due to a TokenBudgetExceeded error or search API failure; recover the run by loading the previous snapshot, applying config fixes, and resuming from that node.
mcleanTRun the AutoReview pipeline locally via Claude Code CLI to generate scientific review papers.
mcleanTScaffold and wire a new DAG node into the AutoReview pipeline.
mahmoud20138Analyze skill usage data and generate optimization recommendations.
openeuler-mirrorCorrelate logs, trace spans, and dependency graphs to output structured JSON root cause reports.
