Skip to content

Wizard and Auto-detect states SE4 for Nordpool, despite SE3 selected #20

Wizard and Auto-detect states SE4 for Nordpool, despite SE3 selected

Wizard and Auto-detect states SE4 for Nordpool, despite SE3 selected #20

Workflow file for this run

# Issue Triage Bot
#
# TRIGGERS:
# issues.opened/reopened → classify, label, and respond to every new issue
# workflow_dispatch → manually re-run triage on any issue
#
# Uses anthropics/claude-code-action@v1 (official Anthropic action).
# Claude reads CLAUDE.md for project context and bot command instructions.
#
# Pipeline overview:
# Stage 1 (auto) — triage-new-issue : classify, label, gather info
# Stage 2 (manual) — @claude-bot confirm : read source and confirm/deny bug
# Stage 3 (manual) — @claude-bot fix this : implement fix, open draft PR
name: Issue Triage
on:
issues:
types: [opened, reopened]
workflow_dispatch:
inputs:
issue_number:
description: "Issue number to triage"
required: true
type: number
jobs:
triage:
name: Triage New Issue
runs-on: ubuntu-latest
permissions:
contents: read
issues: write
steps:
- uses: actions/checkout@v4
- uses: actions/create-github-app-token@v1
id: app-token
with:
app-id: ${{ secrets.CLAUDE_REVIEWER_APP_ID }}
private-key: ${{ secrets.CLAUDE_REVIEWER_PRIVATE_KEY }}
- uses: anthropics/claude-code-action@v1
with:
anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
github_token: ${{ steps.app-token.outputs.token }}
claude_args: "--max-turns 15"
prompt: |
Triage issue #${{ github.event.issue.number || github.event.inputs.issue_number }} in the johanzander/bess-manager repository.
Read docs/agents/rules.md (hard constraints) and docs/agents/architecture.md (codebase layout).
Run: gh issue view ${{ github.event.issue.number || github.event.inputs.issue_number }} --json title,body,labels,comments
Your tasks (in order):
1. Classify the issue: bug, question, feature-request, or other.
2. Apply labels with: gh issue edit <number> --add-label <label>
Available labels: bug, question, feature-request, needs-debug-log, bot-analyzed
3. Post a comment:
- For a potential bug with a debug log already attached: post a focused initial
analysis. Use hedged language ("the log suggests...", "this may indicate...").
Do NOT claim to have confirmed the bug — a separate confirmation step reads
the actual source files. End with:
"A code-level confirmation will follow. To trigger it: `@claude-bot confirm`"
- For a potential bug without a debug log: ask the user to provide one
(System Health → Download Debug Report), label needs-debug-log.
- For a question or feature request: respond helpfully.
- For unclear reports: ask specific clarifying questions.
IMPORTANT: Do NOT read source files to confirm bugs. Do NOT suggest fixes.
Focus only on classification, labeling, and information gathering.