菜单
Use when user provides Jira issue URLs or mentions Jira tickets - fetches issue details and comments from Jira Cloud using local jira tool, outputs AI-optimized markdown for context gathering
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Review a pull request applying the user's Java, Spock, and architecture preferences on top of standard code review. Use when the user asks to review a PR, review code changes, or mentions "my-review". Also use when the user pastes a GitHub PR URL and asks for feedback.
Read and address review comments on a GitHub pull request. Use when the user asks to address PR comments, handle review feedback, respond to PR reviewers, or mentions "address comments" or "pr comments". Also use when the user has just opened a PR and wants to process incoming review feedback. Works with a PR number/URL argument or auto-detects the PR for the current branch.
Reflect on the current conversation session to extract reusable learnings for CLAUDE.md files and to propose improvements to skills that were actually used or invoked during the session. Use when the user asks to reflect on the session, extract learnings, capture lessons, update CLAUDE.md, improve a skill based on this conversation, or invokes /reflect. Optionally focus on a specific area mentioned by the user.
Use to drive the user's cmux terminal multiplexer — read output from other panes, send input/keys into other terminals, spawn new workspaces/splits/panes, control browser surfaces (Playwright-like), and post desktop notifications. Trigger eagerly whenever the user references work happening in another terminal, pane, workspace, or browser surface ("the other terminal", "my running server", "the pane next to this one", "in my browser", "tell me when X finishes"), or asks Claude to open/split/spawn a terminal or browser. Also trigger when the user says "cmux", "send this to…", "look at what's running in…", "notify me when…", or wants Claude to drive a webpage they have open.
Use when the user asks to implement a feature, add a class or method, fix a bug, refactor code, add test coverage, or run autonomously to drive work forward. Supports explicit phase selection via the first argument (red | green | refactor | forever) and infers the phase from conversation and test state when no phase is given. With no arguments at all, defaults to forever (autonomous loop). Do NOT use for code review, CI/CD setup, testing questions, infrastructure, or documentation tasks.
Resolve security vulnerability Jira tickets (typically SECCOMP-*) by upgrading the indicated package, runtime, or base image everywhere it appears in the repository, then creating a branch, commit, and pull request. Use this skill whenever the user asks to "bump", "fix a vuln", "address a SECCOMP ticket", "upgrade a base image", or mentions a ticket ID that looks like a security remediation — even if they don't explicitly say "bump-fix". Also use when the user pastes a Jira URL to a vulnerability ticket and asks to fix it.
| name | jira-tool |
| description | Use when user provides Jira issue URLs or mentions Jira tickets - fetches issue details and comments from Jira Cloud using local jira tool, outputs AI-optimized markdown for context gathering |
Core principle: When users mention Jira issues or provide Jira URLs, immediately run jira <url-or-key>. Don't explore code, don't explain internals, just use the tool.
The jira tool is a Python CLI that fetches Jira Cloud issues (description, comments, metadata) and outputs AI-optimized markdown to stdout.
IMMEDIATE ACTION REQUIRED:
jira "<url or key>"That's it. No analysis, no explanation, no code reading. If jira tool doesn't exists, just tell the user to install it and make use it's in the PATH.
Use jira tool when:
https://*.atlassian.net/browse/KEY-123Don't use when:
| Task | Command |
|---|---|
| Fetch by issue key | jira PROJ-123 |
| Fetch by URL | jira "https://company.atlassian.net/browse/PROJ-123" |
| Save to file | jira PROJ-123 > context.md |
Before using the tool, verify:
# 1. Configuration exists
ls .env
If missing:
.env not found → Tell user to create it with ATLASSIAN_EMAIL, ATLASSIAN_TOKEN, ATLASSIAN_DOMAIN# Direct issue key
jira SP-1234
# Full URL (with or without query params)
jira "https://company.atlassian.net/browse/TICKET-123"
jira "https://company.atlassian.net/browse/TICKET-123?filter=all"
Gather context before implementation:
# Fetch requirements
jira "https://company.atlassian.net/browse/FEATURE-456" > requirements.md
# Read and implement
cat requirements.md
# [implement based on requirements]
Read issue during debugging:
# Fetch bug report
jira BUG-789
# Use output to understand the issue
# [debug based on description and comments]
The tool outputs markdown with:
Structure:
# [KEY-123] Issue Title
**Issue Key:** KEY-123
**Status:** In Progress
**Priority:** High
...
---
## Original Description
[Description content in markdown]
---
## Comments (N total)
### Comment by Author on YYYY-MM-DD HH:MM UTC
[Comment content]
| Error | Cause | Solution |
|---|---|---|
ATLASSIAN_EMAIL not found | Missing .env | Create .env with required vars |
Invalid credentials | Wrong API token | Check ATLASSIAN_TOKEN in .env |
Issue KEY-123 not found | No access or doesn't exist | Verify key and permissions |
Failed to connect to Jira | Network or domain issue | Check ATLASSIAN_DOMAIN in .env |
Invalid issue key | Wrong format | Use format: PROJ-123 |
All errors go to stderr - stdout remains clean for piping.
# DON'T: Spend time reading jira.py source
Read jira to understand API calls...
# DO: Just use it
jira PROJ-123
# DON'T: Fetch HTML page
WebFetch https://company.atlassian.net/browse/PROJ-123
# DO: Use jira tool for structured data
jira PROJ-123
# DON'T: Breaks on query params
jira https://site.atlassian.net/browse/KEY-123?filter=all
# DO: Quote URLs
jira "https://site.atlassian.net/browse/KEY-123?filter=all"