v0.4.0

Changed

• The search_issues tool description now names the clauses that satisfy the
  bounding check and states that scoping to a project is not required — the
  previous "Requires bounded query" wording read as a project-scope
  requirement
• UNBOUNDED_QUERY errors now list the accepted bounding keywords

Added

• Optional per-config timeout field in JIRA_CONFIG_JSON (seconds,
  defaults to 30.0) controlling the HTTP timeout for that instance; a
  non-numeric or non-positive value is rejected at startup
• list_configs now reports each config's effective timeout

Documentation

• Document the HTTP timeout: what the budget covers, why it does not bound
  a large attachment download, and which error codes a timeout surfaces as
• Document that list_configs never returns an error and reports
  {"configs": []} rather than CONFIG_NOT_FOUND for an empty roster
• Note that download_attachment buffers the whole attachment in memory
• Document that JQL values which are reserved words must be quoted
  (project = "ON"), a first-query trap for short project keys

v0.3.0

Changed

• search_issues responses are now field-driven: each issue contains key, url, and only the requested fields (previously a fixed set of keys)
• ADF-to-text conversion joins adjacent inline nodes without inserting spaces (formatting marks no longer split words) and separates block nodes with blank lines instead of single spaces
• key is no longer selectable in search_issues fields — it is always returned, and requesting it now returns VALIDATION_ERROR
• The package installs as src rather than exposing jira, tools, and utils as top-level modules, so python -m src works outside the project directory

Fixed

• JQL bound checks match bounding keywords as whole words and ignore matches inside quoted strings and the ORDER BY clause
• Attachment filename sanitization strips Windows-style path separators and rejects dot-only names (., ..), falling back to attachment
• Attachment downloads report file-write failures as DOWNLOAD_FAILED instead of raising
• search_issues derives is_last from next_page_token when Jira omits isLast, so clients no longer stop paginating while a page remains

v0.2.0

Changed

• Reduced MCP tool schema token usage by ~50%
• Refactored issue creation to use a params object and improved ADF text conversion

Removed

• Deprecated start_at parameter from search_issues (use token-based pagination instead)

v0.1.0

Added

• Initial release
• list_configs tool for listing available Jira configurations
• search_issues tool for JQL-based issue search
• get_issue tool for retrieving issue details with comments and attachments
• create_issue tool for creating new issues
• download_attachment tool for downloading attachments
• Multi-configuration support via JIRA_CONFIG_JSON environment variable
• Input validation for issue keys, project keys, and JQL queries
• ADF (Atlassian Document Format) conversion utilities
• Structured error responses with error codes