Claude-skill-registry check-pr-errors
Check logs for errors and warnings in a specific PR. Use when debugging PR issues, investigating failed code reviews, or checking for errors in a pull request.
install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/check-pr-errors" ~/.claude/skills/majiayu000-claude-skill-registry-check-pr-errors && rm -rf "$T"
manifest:
skills/data/check-pr-errors/SKILL.mdsource content
Check PR Errors
Check logs for errors and warnings in a specific PR.
Usage
Run the check-pr-errors CLI script with the provided arguments:
npx ts-node scripts/check-pr-errors.cli.ts $ARGUMENTS
Arguments
(required): The PR number to checkprNumber
(required): The organization IDorgId
Options
: Number of days to search back (default: 7)--days=N
: Max number of logs to return (default: 100)--limit=N
: Show all log levels (debug, info, warn, error). Default: only error + warn--all
: Also search in legacy collection (observability_logs)--legacy
: Path to .env file (e.g.,--env=PATH
)--env=.env.prod
Examples
# Check errors for PR #723 in production /check-pr-errors 723 97442318-9d2a-496b-a0d2-b45fb --env=.env.prod # Check all logs (not just errors) for PR #701 /check-pr-errors 701 97442318-9d2a-496b-a0d2-b45fb --all --env=.env.prod # Check with more logs and legacy collection /check-pr-errors 701 97442318-9d2a-496b-a0d2-b45fb --limit=500 --legacy --env=.env.prod
What it checks
-
Log errors: Searches
(and optionallyobservability_logs_ts
) for:observability_logs- Logs with level
orerrorwarn - Logs with
presentattributes.error - Logs with messages containing error-related keywords
- Logs with level
-
Telemetry errors: Checks
for failed operations related to the PRobservability_telemetry
Output
The script outputs:
- List of errors/warnings with timestamps, components, and messages
- Stack traces when available
- Failed telemetry operations
- Summary of findings
How to Respond
- Find the root cause of any errors.
- Look for anomalies.
- Pay close attention to skipped reviews.
- Determine the final status (success, error, or skipped). Keep using the command with
if needed.--limit=XX