Multi-Agent MCP Server with tool integrations for GitHub, Slack, and Jira. Provides a REST API and web UI for managing intelligent agents and their tool configurations.
Critical failures across all dimensions. Tool definitions exist only as static JSON objects in tools.ts with name and description fields, but NO input schemas, parameter types, or output schemas are visible in the provided source. Descriptions are minimal (10-50 chars) and lack actionable context. No parameter documentation, no error handling guidance, no security annotations. Tools are inferred from a static array rather than explicitly registered via MCP SDK. This is a REST API wrapper, not a proper MCP server implementation.
Create a comment on a GitHub Pull Request
Get details of a GitHub Pull Request
List GitHub repositories for the authenticated user
Trigger a GitHub Actions workflow
Add a comment to a Jira issue
Create a new Jira issue
Get details of a specific Jira issue
No input schemas visible for any tool. Tools defined as static JSON objects with only 'name', 'description', 'enabled' fields. No parameter types, constraints, or validation rules documented.
Descriptions are 10-50 characters and lack actionable context. E.g., 'List GitHub repositories for the authenticated user' does not explain when to use this vs other tools, what fields are returned, or prerequisites. Below 20-char minimum for scoring.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 33 | <=2025-11-25 | v2 |
Search Jira issues using JQL
Transition a Jira issue to a new status
Update an existing Jira issue
Add a reaction to a Slack message
Get message history from a Slack channel
List available Slack channels
Send a message to a Slack channel
Update an existing Slack message
No parameter documentation. Tools like 'github_create_pr_comment' and 'jira_create_issue' require parameters (PR ID, comment text, issue type, etc.) but none are documented. LLMs cannot infer what parameters to pass.
No error handling or recovery guidance. Tools do not document what errors can occur, whether they are retryable, or what the LLM should do next. No actionable error messages visible.
No security annotations or permission gates. Destructive tools (github_trigger_workflow, slack_send_message, jira_create_issue) lack readOnlyHint/destructiveHint annotations. No audit trail or permission checks visible.