Refactor high-complexity React components in Dify frontend. Use when `pnpm analyze-component...
npx skills add mcncl/skill-buildkite --skill "log-retrieval"
Install specific skill from multi-skill repository
# Description
|
# SKILL.md
name: log-retrieval
description: |
Retrieves and displays job logs from Buildkite. Use when user asks:
- "Show me the logs"
- "What did the build output?"
- "Get the logs from job X"
- "Search the logs for Y"
- "What happened in the test step?"
- /buildkite:logs
Note: For debugging failures, use build-debugging instead. Use this skill for viewing/searching logs without full failure analysis.
Log Retrieval
Retrieve and display job logs from Buildkite builds.
Available MCP Tools
| Tool | Purpose |
|---|---|
buildkite_get_build |
Get build details to find job IDs |
buildkite_read_logs |
Get full log output for a job |
buildkite_search_logs |
Search for patterns within logs |
Input Parsing
Parse from $ARGUMENTS or user's message:
| Input Format | Example |
|---|---|
| Job URL | https://buildkite.com/org/pipe/builds/123#job-uuid |
| Build + job name | build 123, test step |
| Build number | 123 (then ask which job) |
| Description | "the failed job" or "the deploy step" |
Approach
- Identify the job
- If just build given, fetch build and list jobs
- Match job by name/label if specified
-
Default to failed jobs if user asks about errors
-
Fetch logs with
buildkite_read_logs -
Get full output for the identified job
-
Present logs appropriately
- For short logs: show entirely
- For long logs: show relevant section (usually end)
-
Always use code blocks
-
Search if needed with
buildkite_search_logs - When user is looking for specific content
- Return matching lines with context
Output Format
## Logs: "Run Tests" (job abc-123)
โ```
[timestamp] Installing dependencies...
[timestamp] Running test suite...
[timestamp] Error: Connection refused
[timestamp] Test failed: api.test.js
โ```
Showing last 50 lines. Full log is 2,847 lines.
Search Output
When searching:
## Search: "error" in build 456
Found 3 matches:
**Line 234** (test step):
> Error: Cannot connect to database
**Line 567** (deploy step):
> Error: Permission denied
**Line 890** (cleanup step):
> Error: File not found (non-fatal)
When to Use This vs build-debugging
| Scenario | Skill |
|---|---|
| "Why did build fail?" | build-debugging |
| "Show me the test logs" | log-retrieval |
| "Debug this error" | build-debugging |
| "Search for 'timeout' in logs" | log-retrieval |
| "What went wrong?" | build-debugging |
| "What did the deploy output?" | log-retrieval |
Example Interaction
User: Show me the logs from the failed test
1. Get build (from context or ask)
2. Find job with "test" in name that failed
3. Fetch logs with buildkite_read_logs
4. Display relevant portion in code block
# Supported AI Coding Agents
This skill is compatible with the SKILL.md standard and works with all major AI coding agents:
Learn more about the SKILL.md standard and how to use these skills with your preferred AI coding agent.