| name | incident-log-analysis |
| description | Use for analyzing application logs with Python inside the OpenShell sandbox and producing incident reports. |
Incident Log Analysis
Use this skill when the participant asks for application log analysis, incident summaries, error spikes, latency patterns, or remediation recommendations.
Instructions
- Treat
/workshop/data/sample_app.log as the canonical workshop log fixture.
- Copy source data into
/sandbox before writing analysis scripts.
- Write Python scripts under
/sandbox.
- Prefer the Python standard library:
collections, datetime, json, statistics, and re.
- Parse logs into structured records before summarizing.
- Compute concrete evidence:
- counts by level
- counts by service
- counts by status code
- top failing endpoints
- slowest requests
- p95 latency by service when enough samples exist
- first and last error timestamps
- Write two artifacts:
/sandbox/incident_summary.json
/sandbox/incident_report.md
- After writing artifacts, read them back or print a concise preview.
- Do not assume GPU libraries are available.
Report Shape
# Incident Report
## Executive Summary
## Evidence
## Timeline
## Suspected Cause
## Recommended Next Actions