/**
* MCP tool: skippr_get_issue
* Gets full details for a specific issue including raw markdown content
*/
import { z } from 'zod';
import { readIssueFile } from '../utils/issues-reader.js';
// Input schema for the tool
export const GetIssueInputSchema = z.object({
projectId: z.string().describe('Project identifier'),
reviewId: z.string().uuid().describe('Review ID'),
issueId: z.string().uuid().describe('Issue ID'),
});
// Output schema for the tool
export const GetIssueOutputSchema = z.object({
id: z.string().uuid(),
reviewId: z.string().uuid(),
title: z.string(),
severity: z.string(),
resolved: z.boolean(),
category: z.string().optional(),
elementMetadata: z
.object({
selector: z.string(),
bounding_box: z.array(z.number()).optional(),
})
.optional(),
markdown: z.string().describe('Raw markdown content for Claude to read'),
});
export type GetIssueInput = z.infer<typeof GetIssueInputSchema>;
export type GetIssueOutput = z.infer<typeof GetIssueOutputSchema>;
/**
* Get full issue details including raw markdown content
* Claude will read the markdown naturally to understand:
* - Details section
* - Agent Prompt section (actionable instructions)
* - Ticket section (optional, rich user story format)
*/
export async function getIssue(input: GetIssueInput): Promise<GetIssueOutput> {
// Validate input
const validated = GetIssueInputSchema.parse(input);
const { projectId, reviewId, issueId } = validated;
// Read and parse the issue file
const { frontmatter, markdown } = await readIssueFile(projectId, reviewId, issueId);
return {
id: frontmatter.id,
reviewId: frontmatter.reviewId,
title: frontmatter.title,
severity: frontmatter.severity,
resolved: frontmatter.resolved,
category: frontmatter.category,
elementMetadata: frontmatter.elementMetadata,
markdown, // Raw markdown for Claude to read
};
}
