v2.0.1

Listing metadata for MCP directories. No functional changes.

• Adds the ISC LICENSE file (package.json already declared ISC).
• Adds mcpName and server.json for the official MCP Registry.
• Adds glama.json for Glama.

v2.0.0

Every tool is now tested automatically against Jira Server/Data Center 7.13, 8.20, 9.12 (also under /jira) and 10.3, and CI runs on Windows, macOS and Linux (Node 18–24).

⚠️ Breaking changes

• jira_get_issue returns customFields as [{ id, name, value }] rows instead of an object. Comments, worklogs and attachments are opt-in via include (#1).
• jira_search returns a compact summary by default. fields now adds to the defaults; use jira_get_issue for full details.

New

• jira_get_edit_meta: which fields can be edited on an issue. Custom fields can be written by unique display name (#1, thanks @PHMT-CW).
• jira_get_issue_images and jira_upload_attachment (thanks OptiMiser-LLC and @rainmakerho).
• customFields on create and update (thanks @rainmakerho).
• Credentials from the OS keychain as an optional fallback to env vars (thanks @rainmakerho).

Fixed

• jira_create_link failed even though the link was created (#2).
• jira_get_issue_types / jira_get_create_meta on Jira 9 and 10.
• jira_search crash when fields is passed; jira_update_sprint needing every field.
• Writes are no longer retried after timeouts, which could create duplicates.
• After a failed login the server stops retrying, so agent retries can't trigger a Jira CAPTCHA lockout.
• JIRA_PROJECTS_FILTER and JIRA_SSL_VERIFY=false now work.
• Path arguments are URL-encoded; uploads reject network paths and files over 100 MiB.
• Clear errors for redirects, a wrong JIRA_BASE_URL or context path, HTML error pages and real 403 reasons.

Published from GitHub Actions with npm provenance.