jira-cookie-automation

name

jira-cookie-automation

description

Automate Jira work with a live browser session cookie, automatically bootstrapping `JIRA_COOKIE` from an existing Chrome DevTools MCP Jira browser session when available. Use when Codex needs to create, file, update, assign, comment on, link, transition, search, sprint-assign, verify, or batch-process Jira issues, especially when a repo has planning CSV/JSON files or when the user wants Jira website actions made repeatable from any workspace.

Jira Cookie Automation

Core Rule

Prefer cookie-backed Jira REST automation over manual browser navigation. Use manual Jira UI only when REST is blocked, the user explicitly asks for website navigation, or an operation cannot be represented safely through the API.

Before asking the user to paste JIRA_COOKIE, try to extract the Cookie request header from an existing authenticated Jira page in Chrome DevTools MCP. Never store Jira cookies in files, logs, skill resources, commits, or generated artifacts.

Cookie Bootstrap Order

1. If Chrome DevTools MCP tools are available, use references/mcp-cookie-bootstrap.md and try to read the Cookie header from an authenticated Jira network request.
2. If MCP extraction succeeds, set JIRA_BASE_URL from the request origin and use that Cookie header as JIRA_COOKIE for the CLI command.
3. If MCP extraction fails because no Jira page or request is available, ask the user to open Jira and trigger any authenticated REST request, such as /rest/api/3/myself.
4. Ask the user to paste a cookie manually only when MCP is unavailable or cannot expose request headers.

Quick Start

Use the bundled CLI from this skill directory:

$env:JIRA_BASE_URL = 'https://your-site.atlassian.net'
$env:JIRA_COOKIE = 'paste the full Cookie request header here'
python scripts\jira_cookie.py whoami

Run write operations as dry-runs first:

python scripts\jira_cookie.py create --dry-run --project APP --type Task --summary "Build shared shell" --description "Implement:`n- Build the route shell"

Then run the same command without --dry-run only after the payload is correct.

Workflow

1. Identify the Jira base URL, project key, issue type, board ID, labels, priority names, and assignee names or account IDs from the repo, user message, or Jira itself.
2. Bootstrap JIRA_COOKIE from Chrome DevTools MCP when available. Fall back to manual cookie setup only if needed.
3. Run whoami to verify the cookie and base URL.
4. Use search, get, user-search, and transitions to discover current Jira state before writing.
5. For writes, run with --dry-run first and inspect the JSON payload.
6. Apply small batches. For issue creation, keep batches to 3 to 5 unless the user explicitly wants a larger automated run.
7. Verify by reading Jira back after each batch. Prefer the CLI's --verify flag where available.
8. Keep a local progress file or write-back CSV updated after successful batches.

Common Commands

Authentication check:

python scripts\jira_cookie.py whoami

Search issues:

python scripts\jira_cookie.py search --jql "project = APP ORDER BY created DESC" --fields summary,status,assignee,priority

Create an issue:

python scripts\jira_cookie.py create --project APP --type Task --summary "Implement profile shell" --description-file issue.txt --assignee "Jordan Hale" --priority High --label app-v2 --verify

Update summary, description, assignee, priority, or labels:

python scripts\jira_cookie.py update APP-123 --summary "Implement profile shell route" --assignee "Jordan Hale" --priority Normal --verify

Assign an issue:

python scripts\jira_cookie.py assign APP-123 --assignee "Riley Chen" --verify

Add a dependency link:

python scripts\jira_cookie.py link --type Blocks --inward APP-124 --outward APP-123

This means Jira stores a Blocks link where APP-124 is blocked by APP-123.

Transition an issue:

python scripts\jira_cookie.py transition APP-123 --to Done --comment "Scope covered by APP-122."

Assign issues to a sprint:

python scripts\jira_cookie.py assign-sprint --board-id 42 --sprint "App Sprint 6" --create-missing --issues APP-123 APP-124

Batch-create from CSV:

python scripts\jira_cookie.py batch-create --csv jira-ui-create-queue.csv --project APP --type Task --summary-col Summary --description-col Description --assignee-col Assignee --priority-col Priority --label app-v2 --limit 5 --write-back --verify

See references/batch-formats.md for CSV and JSON plan patterns.

Operating Rules

• Treat JIRA_COOKIE as a short-lived session secret. If a request returns 401, 403, CAPTCHA, or unexpected HTML, ask the user to refresh the cookie.
• Prefer an MCP-extracted cookie over asking the user to copy/paste one.
• Do not invent assignee account IDs. Resolve names with user-search or use IDs already present in trusted local files.
• Do not create duplicate issues. Search by summary, project, labels, or known issue keys before creating.
• Do not silently close or transition issues. Comment first when closing duplicates or retargeting scope.
• Use Jira read-back verification for summary, assignee, priority, labels, status, sprint, and links after writes.
• Keep project-specific constants out of the skill. Use environment variables, command flags, or repo-local config/docs.
• Keep generated logs free of cookie values and request headers.

Resources

• scripts/jira_cookie.py: reusable Jira REST CLI backed by JIRA_COOKIE.
• references/mcp-cookie-bootstrap.md: Chrome DevTools MCP flow for extracting the Cookie header from an existing Jira browser session.
• references/cookie-auth.md: safe cookie acquisition and environment setup.
• references/batch-formats.md: reusable CSV/JSON shapes for filing, assignment, linking, sprint work, and resumable runs.
• references/manual-ui-fallback.md: condensed website-navigation fallback for the Jira create-issue UI.