run-evaluation
Evaluate the latest Universal Agent run for errors, bottlenecks, and opportunities for improvement. This skill should be used after an agent run completes to perform a critical assessment by analyzing the run.log file, session directory output, and Logfire traces. Use when the user wants to debug issues, understand performance problems, identify exceptions, check if the agent stayed on a happy path, or get recommendations for improving agent behavior.
npx skills add majiayu000/claude-skill-registry --skill run-evaluation --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
# Run Evaluation Skill Perform a comprehensive post-mortem analysis of the latest Universal Agent run. ## Workflow ### Step 1: Identify the Latest Session Find the most recent session directory: ```bash ls -lt /home/kjdragan/lrepos/universal_agent/AGENT_RUN_WORKSPACES/ | grep session_ | head -1 ``` Extract the session path (e.g., `/home/kjdragan/lrepos/universal_agent/AGENT_RUN_WORKSPACES/session_20260115_094820`). ### Step 2: Read the Run Log Load the full run log for context: ```bash cat {session_dir}/run.log ``` This contains the complete terminal output including: - Tool calls and responses - Error messages - Timing information - Agent decisions ### Step 3: Extract Key Metrics from run.log Parse the log for: - **Tool call count**: Count occurrences of `🔧 [` - **Error indicators**: Search for `Error`, `Failed`, `Exception`, `❌` - **Timing**: Look at `+Xs` timestamps for latency - **Retries/deduplication**: Search for `Idempotent`, `retry`, `deduped` ### Step 4: Query Logfire for Trace Analysis Use the Logfire MCP tools to analyze the run. Get the trace_id from the run.log (appears near the top). **Key queries:** 1. **Find all exceptions in the run:** ```sql SELECT start_timesta
- Workflow
- Step 1: Identify the Latest Session
- Step 2: Read the Run Log
- Step 3: Extract Key Metrics from run.log
- Step 4: Query Logfire for Trace Analysis
- Step 5: Analyze Session Artifacts
- Step 6: Generate Evaluation Report
- Evaluation Report Template
- Evaluation Criteria
- Happy Path Indicators (✅)
- Warning Indicators (⚠️)
- Critical Indicators (❌)
- Output
ls -lt /home/kjdragan/lrepos/universal_agent/AGENT_RUN_WORKSPACES/ | grep session_ | head -1
cat {session_dir}/run.log
find {session_dir} -type f -name "*.md" -o -name "*.json" -o -name "*.html" | head -30What does the run-evaluation skill do?
Evaluate the latest Universal Agent run for errors, bottlenecks, and opportunities for improvement. This skill should be used after an agent run completes to perform a critical assessment by analyzing the run.log file, session directory output, and Logfire traces. Use when the user wants to debug issues, understand performance problems, identify exceptions, check if the agent stayed on a happy path, or get recommendations for improving agent behavior.
How do I install it?
Run `npx skills add majiayu000/claude-skill-registry --skill run-evaluation --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From majiayu000/claude-skill-registry, a repository with 534 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.