redmine-issue-report

name: redmine-issue-report description: Use when querying Redmine issues, generating issue reports, or creating issue summary markdown files. Triggers on "redmine", "issue report", "issue list", "bug list".

Redmine Issue Report

Overview

Query Redmine issues via REST API and generate a formatted markdown report with issue details and descriptions.

Trigger

Use when user:

• Wants to query or list Redmine issues
• Needs a summary report of issues (by assignee, status, priority, etc.)
• Asks to generate a markdown report from Redmine

Skill Directory

SKILL_DIR=".claude/skills/redmine-issue-report"

Authentication Flow

Credentials are stored in a .env file inside the skill directory ($SKILL_DIR/.env).

Step 1: Check for credentials

Read $SKILL_DIR/.env and verify REDMINE_URL, REDMINE_API_KEY, and REDMINE_PROJECT are set.

• If .env exists and all values are set → proceed to Workflow
• If .env exists but values are empty → ask user to fill in, go to Step 3
• If .env doesn't exist → go to Step 2

Step 2: Create .env file (if doesn't exist)

Create $SKILL_DIR/.env with template:

REDMINE_URL=
REDMINE_API_KEY=
REDMINE_PROJECT=
REDMINE_USER=

Ensure .env is in .gitignore to prevent credential leaks.

Then display this message to the user:

I've created a .env file at .claude/skills/redmine-issue-report/.env
Please edit the file and fill in your Redmine credentials:

REDMINE_URL=https://redmine.vivotek.tw
REDMINE_API_KEY=your_api_key_here
REDMINE_PROJECT=vsaas
REDMINE_USER=your_login_name

After updating the file, let me know and I'll continue.

Step 3: Load credentials

Read values from .env and use them for all subsequent API calls. Never log or echo the API key.

Filter Parameters

These are provided by the user at query time (not stored in .env):

Workflow

Step 1: Resolve User ID (if filtering by assignee)

CRITICAL: Redmine API assigned_to_id only accepts numeric IDs, NOT login names.

Do NOT attempt assigned_to_id=login_name — it returns 0 results silently.

Strategy: Fetch first page of issues, find the target assignee's numeric ID from the assigned_to field in any matching issue, then use that ID for the full query.

GET {REDMINE_URL}/projects/{project}/issues.json?key={api_key}&limit=100&status_id=*&sort=updated_on:desc

From results, find: "assigned_to": {"id": 34, "name": "alive.kuo"} → use assigned_to_id=34

Step 2: Fetch All Issues

Once you have the numeric user ID, query with server-side filtering:

GET {REDMINE_URL}/issues.json?key={api_key}&assigned_to_id={user_id}&limit=100&status_id={status}&sort=id:desc

Important parameters:

• status_id=* → all statuses (open + closed)
• status_id=open → only open issues
• status_id=closed → only closed issues
• limit=100 → max per page
• offset=N → for pagination (if total_count > 100)

Check total_count in response. If more than 100, paginate with offset=100, offset=200, etc.

Step 3: Fetch Issue Details

For each issue, fetch full details including description:

GET {REDMINE_URL}/issues/{issue_id}.json?key={api_key}

The description field contains the full issue body with formatting.

Performance tip: Fetch multiple issue details in parallel using the Agent tool or parallel WebFetch calls.

Step 4: Generate Markdown Report

CRITICAL: You MUST write the report to a .md file using the Write tool. Do NOT only display results in conversation.

After writing the file, provide a brief summary to the user in conversation.

Output filename: redmine_issue_{YYYY-MM-DD}.md (e.g., redmine_issue_2026-03-06.md)

# {Project} Issue Report

**Generated:** {date}
**Filter:** Assignee: {name} | Status: {status}
**Total Issues:** {count}

## Summary

| ID | Subject | Status | Priority | Assigned To |
|----|---------|--------|----------|-------------|
| #ID | Subject text | Status | Priority | Name |
| ... | ... | ... | ... | ... |

## Issue Details

### #{id} - {subject}

- **Status:** {status}
- **Priority:** {priority}
- **Tracker:** {tracker}
- **Assigned To:** {name}
- **Author:** {author}
- **Created:** {date}
- **Updated:** {date}

**Description:**

{Copy the full description from Redmine as-is. Do NOT summarize, paraphrase, or restructure.}

---

(repeat for each issue)

Common Mistakes

WebFetch Limitations

The WebFetch tool processes content through a small AI model. When dealing with large JSON (100+ issues), this model may miss entries. Always:

1. Use server-side filtering (query parameters) instead of client-side filtering
2. Verify total_count matches the number of results you received
3. For assignee queries, resolve user ID first, then let Redmine filter