jira · odyssey4me — скилл для Claude Code: поиск и управление задачами в Jira
Скилл для Claude Code, который работает с задачами Jira через JQL: ищет, создаёт и обновляет тикеты, читает комментарии. Запускается скриптом jira.py, которому нужен доступ к API вашего инстанса.
MIT 139 установок ★ 13 обновлён 2026-10-25
npx skills add odyssey4me/agent-skills --skill jira -a claude-codeЧто делает jira
Внутри Python-скрипт с командами search, issue get, issue create, issue update и issue comment, плюс check для проверки конфигурации.
Поддерживается только Jira Cloud. Аутентификация через переменные окружения JIRA_BASE_URL, JIRA_EMAIL, JIRA_API_TOKEN или через файл ~/.config/agent-skills/jira.yaml. API-токен создаётся на странице Atlassian, скрипт не сбросит пароль и не выдаст сетевой доступ — если агент работает в песочнице, доступ к HTTPS нужно открыть до первого запроса.
Поиск принимает JQL-строку, фильтры по участнику, команде, проекту и лимит --max-results (по умолчанию 50). Создавать задачи можно из файла через --from-file, батчами читать несколько тикетов за один вызов, а --contributors собирает роли reporter, assignee и commenter по всем комментариям. Python-зависимости: requests, keyring, pyyaml, pyadf.
Какие задачи решает
- найти открытые тикеты проекта по JQL
- создать задачу из markdown-файла
- вытащить всех комментаторов из длинного треда
- обновить статус, приоритет и story points
- проверить конфигурацию и связь с Jira одной командой
Примеры запросов агенту
- Найди все открытые тикеты проекта DEMO и покажи сводку
- Создай задачу типа Story с родителем EPIC-123 по описанию из issue.md
- Добавь комментарий к DEMO-123: «Деплой прошёл успешно»
Как установить jira
1Через Skills CLI (Vercel)
npx skills add odyssey4me/agent-skills --skill jira -a claude-code2Вручную: положить файлы
~/.claude/skills/jira/SKILL.mdПапка скилла целиком; для одного проекта — .claude/skills/jira/
1Через Skills CLI (Vercel)
npx skills add odyssey4me/agent-skills --skill jira -a codex2Вручную: положить файлы
.agents/skills/jira/SKILL.mdДля всех проектов — ~/.agents/skills/
1Через Skills CLI (Vercel)
npx skills add odyssey4me/agent-skills --skill jira -a cursor2Вручную: положить файлы
.cursor/skills/jira/SKILL.mdCursor читает и .agents/skills/, и .claude/skills/
1GitHub CLI
gh skill install odyssey4me/agent-skills skills/jira2Вручную: положить файлы
.github/skills/jira/SKILL.md1Через Skills CLI (Vercel)
npx skills add odyssey4me/agent-skills --skill jira -a gemini-cli2Командой платформы
gemini skills install https://github.com/odyssey4me/agent-skills --path skills/jiraВручную: положить файлы
.devin/skills/jira/SKILL.mdСтарый путь .windsurf/skills/ тоже читается
Через Skills CLI (Vercel)
npx skills add odyssey4me/agent-skills --skill jira -a opencodeClaude.ai
Упакуйте папку в zip и загрузите в Settings → Capabilities → Skills
По шагам, с командами для каждой платформы: Как установить скилл в Claude Code: пять способов
Вопросы о jira
Как установить jira в Claude Code?
Выполните npx skills add odyssey4me/agent-skills --skill jira -a claude-code, затем откройте ~/.claude/skills/jira/SKILL.md.
Что делает скилл jira?
Ищет задачи по JQL, создаёт и обновляет тикеты, читает комментарии и дочерние задачи, работает со спринтами и бэклогом через Jira Cloud API.
Какие зависимости нужны для работы?
Python-пакеты requests, keyring, pyyaml, pyadf — ставятся через pip install --user requests keyring pyyaml pyadf. Плюс URL инстанса, email и API-токен Atlassian.
Исходный текст
SKILL.md · MIT
--- name: jira description: Search and manage Jira issues using JQL queries, create/update tickets, and manage workflows. Use when asked to find Jira tickets, check the backlog, manage sprints, track bugs, or work with Atlassian project management. metadata: author: odyssey4me version: "1.3.0" category: project-management tags: "issues, agile, sprints" complexity: standard license: MIT --- # Jira Interact with Jira for issue tracking, search, and workflow management. ## Resolving script paths `SKILL_DIR` in the examples is the directory containing this loaded `SKILL.md`. Set it to that absolute path before running commands; do not assume the agent sets this shell variable. Quote the script path if it contains spaces. ## Installation 1. **Install Python dependencies**: ```bash pip install --user requests keyring pyyaml pyadf ``` 2. **Download the skill** from [Releases](https://github.com/odyssey4me/agent-skills/releases) or use directly from this repository. ## Setup Verification After installation, verify the skill configuration by running: ```bash $SKILL_DIR/scripts/jira.py check ``` This will check: - Python dependencies (requests, keyring, pyyaml, pyadf) - Authentication configuration - Connectivity to Jira If anything is missing, the check command will provide setup instructions. ## Network access Jira API commands require outbound HTTPS to the configured Jira host. Remote operations require outbound network access. If the agent environment restricts it, obtain access through that environment's approval or network controls before the first remote command, reusing authorization already granted. Local help and file processing do not need network access. Skill instructions do not grant network permission. Report DNS, connection, and timeout failures as network errors; do not reset credentials to resolve them. If approved access also fails, stop and report it. Authentication and permission errors require user action. ## Authentication This skill supports Jira Cloud only. Configure Jira authentication using one of these methods: ### Option 1: Environment Variables (Recommended) ```bash export JIRA_BASE_URL="https://yourcompany.atlassian.net" export JIRA_EMAIL="[email protected]" export JIRA_API_TOKEN="your-token" ``` Add these to your `~/.bashrc` or `~/.zshrc` for persistence. ### Option 2: Config File Create `~/.config/agent-skills/jira.yaml`: ```yaml url: https://yourcompany.atlassian.net email: [email protected] token: your-token ``` ### Required Credentials - **URL**: Your Jira instance URL (e.g., `https://yourcompany.atlassian.net`) - **Email**: Your Atlassian account email - **API Token**: Create at https://id.atlassian.com/manage-profile/security/api-tokens ## Configuration Defaults Optionally configure defaults (JQL scope, fields, custom fields, project defaults) in `~/.config/agent-skills/jira.yaml`. CLI arguments always override config defaults. See [configuration.md](references/configuration.md) for the full config format and defaults behavior. ```bash # Show all configuration $SKILL_DIR/scripts/jira.py config show # Show project-specific defaults $SKILL_DIR/scripts/jira.py config show --project DEMO ``` ## Commands See [permissions.md](references/permissions.md) for read/write classification of each command. ### check Verify configuration and connectivity. ```bash $SKILL_DIR/scripts/jira.py check ``` This validates: - Python dependencies are installed - Authentication is configured - Can reach the Jira Cloud API - Automation API access - Custom field mappings ### search Search for issues using JQL (Jira Query Language). ```bash $SKILL_DIR/scripts/jira.py search "project = DEMO AND status = Open" $SKILL_DIR/scripts/jira.py search "assignee = currentUser() ORDER BY updated DESC" --max-results 20 ``` **Arguments:** - `jql`: JQL query string (required unless `--contributor`, `--team`, or `--team-name` is used) - `--contributor`: Search for issues where this user is a contributor (reporter, assignee, or updater). Prefer an accountId or exact email; display names can be ambiguous. - `--project`: Project key to scope a `--contributor` search - `--max-results`: Maximum number of results (default: 50) - `--fields`: Comma-separated list of fields to include - `--team`: Team ID to scope the search; repeatable - `--team-name`: Team name to resolve (shared teams only); repeatable Text output renders summary, type, parent, status, resolution and assignee; use `--json` for other fields. Bulk comment authors: see [examples.md](references/examples.md#collect-comment-authors-in-bulk). **JQL patterns:** See [jql-reference.md](references/jql-reference.md) for Cloud-native JQL patterns (status, dates, fields, ordering, linked issues, hierarchy). ### issue Get, create, update, or comment on issues. ```bash # Get issue details (--fields, --contributors) $SKILL_DIR/scripts/jira.py issue get DEMO-123 $SKILL_DIR/scripts/jira.py issue get DEMO-123 --fields "summary,status,assignee" --contributors # Batch several issues into one search call $SKILL_DIR/scripts/jira.py issue get DEMO-1,DEMO-2,DEMO-3 --contributors # List comments $SKILL_DIR/scripts/jira.py issue comments DEMO-123 # List child issues and subtasks, or only their distinct assignees $SKILL_DIR/scripts/jira.py issue children EPIC-123 $SKILL_DIR/scripts/jira.py issue children EPIC-123 --assignees-only $SKILL_DIR/scripts/jira.py issue children EPIC-1,EPIC-2 --assignees-only # Create (--project, --type, --summary required; --description, --priority, --labels, --assignee, --parent, --set-field, --link, --from-file, --json) $SKILL_DIR/scripts/jira.py issue create --project DEMO --type Task --summary "New task" $SKILL_DIR/scripts/jira.py issue create --project DEMO --type Story --summary "New story" --parent EPIC-123 $SKILL_DIR/scripts/jira.py issue create --from-file issue.md --priority Critical # Update (--summary, --description, --priority, --labels, --assignee, --parent, --remove-parent, --set-field, --link, --unlink, --from-file) $SKILL_DIR/scripts/jira.py issue update DEMO-123 --summary "Updated" --set-field story_points=5 $SKILL_DIR/scripts/jira.py issue update DEMO-123 --parent EPIC-456 $SKILL_DIR/scripts/jira.py issue update DEMO-123 --remove-parent $SKILL_DIR/scripts/jira.py issue update DEMO-123 --link "Blocks:DEMO-456" $SKILL_DIR/scripts/jira.py issue update DEMO-123 --unlink "Blocks:DEMO-456" # Comment (--security-level for private comments) $SKILL_DIR/scripts/jira.py issue comment DEMO-123 "This is a comment" ``` `--contributors` reports each person with the roles they hold: `reporter`, `assignee`, and `commenter`. Every comment is read, so a commenter late in a long thread is still reported. With `--json`, `_contributors` is `[{"name": "Jane Doe", "roles": ["reporter", "commenter"]}]`. A batch reads the `comment` field from the search result, falls back to the full thread when the page is truncated, and uses the default fields (no description, issue links or web links); pass `--fields` to request more. `issue children --assignees-only` prints just the distinct assignee names — the cheap way to get the collaborator set. See [from-file-format.md](references/from-file-format.md) for the `--from-file` markdown format and [examples.md](references/examples.md) for more patterns. ### transitions Manage issue workflow transitions. ```bash # List available transitions $SKILL_DIR/scripts/jira.py transitions list DEMO-123 # Transition issue $SKILL_DIR/scripts/jira.py transitions do DEMO-123 "In Progress" $SKILL_DIR/scripts/jira.py transitions do DEMO-123 "Done" --comment "Completed" # Transition with private comment $SKILL_DIR/scripts/jira.py transitions do DEMO-123 "Done" --comment "Internal resolution notes" --security-level "Internal" ``` ### Additional Commands See [advanced-commands.md](references/advanced-commands.md) for full documentation of these commands: - **config** — view/manage configuration defaults and discover custom field mappings - **fields** — list available fields (global or per project/issue type) - **statuses** — list statuses and status categories - **user** — search users by email/name (returns accountId on Cloud) - **collaboration** — discover cross-team collaboration patterns across issue types - **automations** — list and inspect Jira automation rules (Cloud-only) - **versions** — list project versions (releases) and view release reports ## JQL Reference Common JQL queries and patterns: see [jql-reference.md](references/jql-reference.md). Quick reference — combine with `AND`, `OR`, and `ORDER BY`: ```jql assignee = currentUser() AND statusCategory != Done ORDER BY priority DESC ``` Use `statusCategory` (`"To Do"`, `"In Progress"`, `Done`) for queries that work across projects. ## Examples Common Jira workflows: **Search and view:** ```bash # Find all open tasks assigned to you jira search "assignee = currentUser() AND status = Open" # Get issue details jira issue get DEMO-123 --fields "summary,status,assignee" # List issue comments jira issue comments DEMO-123 ``` **Create and update:** ```bash # Create a new bug jira issue create --project DEMO --type Bug --summary "Login page broken" # Create a story under an epic jira issue create --project DEMO --type Story --summary "Implement feature" --parent EPIC-123 # Set or change an issue's parent (works for any hierarchy level) jira issue update DEMO-123 --parent EPIC-456 # Remove an issue's parent jira issue update DEMO-123 --remove-parent # Update an issue and add a comment jira issue update DEMO-123 --priority High --assignee "[email protected]" jira issue comment DEMO-123 "Assigned to review" ``` **Workflow:** ```bash # List available transitions jira transitions list DEMO-123 # Move issue to Done jira transitions do DEMO-123 "Done" --comment "Completed" ``` ## Model Guidance This skill makes API calls requiring structured input/output. A standard-capability model is recommended.
© odyssey4me · MIT · odyssey4me/agent-skills
Похожие скиллы
Все скиллы по теме «Настройка агента: память, контекст, субагенты» →Сначала — jira от других авторов.