Backlog MCP Server - Integrates Backlog project management with Model Context Protocol
The server has 12 tools with complete input schemas and descriptions present. However, quality varies significantly. Naming is mostly clear with verb prefixes (get_, list_, add_, update_). Descriptions are present but often generic and lack actionable context (e.g., 'Get Backlog space information' doesn't explain WHEN to use it or WHAT structure is returned). Parameter descriptions are present but inconsistent in detail. Many parameters lack type constraints (enums, ranges, formats). No output schemas are documented anywhere in the source. Error handling is not visible in the tool definitions, no recovery guidance, error categorization, or validation rules. The server accepts parameter-based auth (not visible but typical for Backlog clients) rather than using the MCP secret injection pattern. Tools like get_project_issues and get_wiki_page_list accept optional parameters but don't document defaults, which invites ambiguous LLM behavior. Overall, this is a competent but not exceptional toolkit, functional definitions with missing polish on error handling, output schemas, and parameter constraints.
Add a comment to a specific Backlog issue
Get Backlog space information
Get current Backlog user information
Get detailed information about a specific comment in a Backlog issue
Get the count of comments in a specific Backlog issue
Get comments from a specific Backlog issue
No output schemas documented for any tool. LLMs cannot plan downstream chaining or understand what fields are returned. This is a critical blocker for agentic reasoning.
Many optional parameters lack documented defaults. For example, list_recent_projects accepts 'count' and 'order' but doesn't state what happens if omitted. LLMs will guess or omit parameters, causing inconsistent behavior.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-22 | D | 52 | 2026-07-28+ | v2 |
| 2026-03-09 | F | 48 | - | v1 |
Get detailed information about a specific Backlog issue
Get issues from a Backlog project
Get detailed information about a specific Wiki page
Get a list of Wiki pages from Backlog
List recently viewed Backlog projects
Update a Wiki page in Backlog
No error handling or recovery guidance in tool definitions. The source shows a generic try-catch in executeTools but no tool-specific error messages or guidance on what the LLM should do when a call fails (retry? ask user? try alternative tool?).
Descriptions are generic and lack WHEN/WHY context. For example, 'Get current Backlog user information' doesn't explain when to call this vs other tools, what it uniquely provides, or how it's used in typical workflows. This is below the 10 - 1024 character guidance and lacks detail.
No tool annotations (readOnlyHint, destructiveHint, idempotentHint). All tools are manually marked with Risk fields, but the MCP protocol supports annotations via tool metadata. This limits agent guardrails.
Parameters describing IDs lack consistent typing. Some use string (issueIdOrKey, projectIdOrKey, wikiId), others use number (commentId, statusId, assigneeId). The schema type is correct but LLM confusion on whether to pass '123' or 'ABC-123' is invited. Consider aliasing with both _id and _key suffixes or explicit examples.
No pagination documentation for list/get-many tools. get_project_issues and get_issue_comments accept count/offset but no output schema clarifies whether there's a total_count, next_cursor, or has_more field. Large result sets risk context explosion.
Redundant tools not justified. get_issue_comment_count seems redundant if get_issue_comments already returns a total_count. The description should explain the unique use case or consolidate into one tool.