メニュー
Detects duplicate JIRA bugs before creation by comparing new bug summaries against existing issues. Use when QA is logging bugs, mentions potential duplicates, or before creating new JIRA tickets.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Generates n8n workflow JSON files from user prompts for download and import. Use when user wants to create n8n automation, mentions workflow generation, or needs a .json file for n8n import.
Runs automated smoke tests after deployment to validate critical paths and UI readiness. Use after deployments, before releases, or when verifying environment health.
Validates UI text content for spelling, grammar, formatting, and consistency issues. Use when testing web pages, reviewing UI copy, checking for content errors, or validating labels and messages.
Processes failed test executions, screenshots, and logs to generate comprehensive bug reports and automatically create JIRA tickets. Includes root cause analysis and environment details.
Automatically generates valid, invalid, and boundary test data based on field names and validation rules. Use when the user provides field specifications and needs a comprehensive data set for testing forms, APIs, or database constraints.
Analyzes requirement documents and existing test cases (Excel/JIRA) to identify coverage gaps, duplicates, and missing scenarios. Generates a gap analysis report mapping requirements to test coverage status.
| name | detecting-duplicate-bugs |
| description | Detects duplicate JIRA bugs before creation by comparing new bug summaries against existing issues. Use when QA is logging bugs, mentions potential duplicates, or before creating new JIRA tickets. |
Prevents JIRA clutter by identifying duplicate or similar bugs before ticket creation.
Gather from user input:
Construct JQL query to find potential duplicates:
project = * AND (
summary ~ "keyword1 keyword2" OR
description ~ "keyword1 keyword2" OR
text ~ "error message"
) AND status != Closed ORDER BY updated DESC
Search strategies:
For each candidate bug, compare:
| Field | Weight | Method |
|---|---|---|
| Summary | 40% | Cosine similarity or keyword overlap |
| Description | 40% | Semantic similarity, error pattern match |
| Component | 10% | Exact match bonus |
| Steps to Reproduce | 10% | Sequence similarity |
Similarity scoring:
Response format:
Duplicate Check Results:
━━━━━━━━━━━━━━━━━━━━━━━
🔍 Analyzed: [X] existing bugs
✅ HIGH MATCH (>85%):
• BUG-245: "[Summary]" (92% similar)
→ URL: https://jira.company.com/browse/BUG-245
⚠️ POSSIBLE MATCH (70-85%):
• BUG-189: "[Summary]" (78% similar)
→ URL: https://jira.company.com/browse/BUG-189
✨ Recommendation: [Create New / Review BUG-245 / Add Comment to BUG-189]
# Search issues
GET /rest/api/2/search?jql={query}&fields=summary,description,status,key
# Get issue details
GET /rest/api/2/issue/{issueKey}
JIRA_API_TOKENJIRA_BASE_URL (e.g., https://company.atlassian.net)Use the helper script for similarity calculation:
python .agent/skills/detecting-duplicate-bugs/scripts/similarity_checker.py \
--new-bug "bug_summary.txt" \
--candidates "candidate_bugs.json" \
--threshold 0.7
User: "Login button not working on mobile safari"
→ Search JQL: project = * AND (summary ~ "login button" OR summary ~ "mobile safari")
→ Compare with top 20 results
→ Output: "Similar bug found: BUG-178 (85% match) - Login failure on iOS browsers"
Set in environment or .env file:
JIRA_BASE_URL=https://yourcompany.atlassian.net
JIRA_API_TOKEN=your_api_token_here
JIRA_PROJECT_KEYS=PROJ,QA,BUG
SIMILARITY_THRESHOLD=0.75