fix(portals): depth-track div extraction so nested job descriptions aren't truncated (#204)

The jobindex and linkedin detail parsers matched description containers with a non-greedy regex that stops at the first inner </div>, so any posting whose description contains nested divs was silently truncated (jobindex dropped later sections; linkedin dropped everything after the first block). Replaces the regex with a depth-tracked extractDivContent scanner that walks div open/close markers to the matching close. Verified: truncation bug reproduced against real markup fixtures, depth arithmetic correct (no off-by-one/infinite-loop), 28 tests pass network-free, no regression on non-nested divs. Malformed-HTML over-grabs rather than truncates - the safer failure, cleaned by downstream stripTags/decode.

By @oscarbol09.

.agents/skills/jobindex-search/cli/src/commands/detail.ts

@@ -1,6 +1,6 @@
 import { defineCommand, option } from "@bunli/core"
 import { z } from "zod"
-import { htmlFetch, writeError } from "../helpers.js"
+import { htmlFetch, writeError, extractDivContent } from "../helpers.js"
 
 const BASE_URL = "https://www.jobindex.dk"
 
@@ -180,9 +180,9 @@ function parseDetailPage(html: string, url: string, id: string): DetailResult {
   let description: string | null = null
 
   // Try job-text class first
-  const jobTextMatch = html.match(/class="job-text"[^>]*>([\s\S]*?)<\/div>\s*(?:<div|<\/div>)/i)
-  if (jobTextMatch) {
-    description = decodeHtmlEntities(stripTags(jobTextMatch[1])).replace(/\s+/g, " ").trim() || null
+  const jobTextHtml = extractDivContent(html, "job-text")
+  if (jobTextHtml) {
+    description = decodeHtmlEntities(stripTags(jobTextHtml)).replace(/\s+/g, " ").trim() || null
   }
 
   // Fallback: try og:description meta tag for a brief description

.agents/skills/jobindex-search/cli/src/helpers.ts

@@ -307,6 +307,33 @@ export function parseJobCards(html: string): JobCard[] {
   return results
 }
 
+export function extractDivContent(html: string, className: string): string | null {
+  const escaped = className.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
+  const openRe = new RegExp(`<div[^>]*class="[^"]*${escaped}[^"]*"[^>]*>`, 'i')
+  const open = openRe.exec(html)
+  if (!open) return null
+
+  let i = open.index + open[0].length
+  let depth = 1
+
+  while (depth > 0 && i < html.length) {
+    const nextOpen = html.indexOf('<div', i)
+    const nextClose = html.indexOf('</div>', i)
+
+    if (nextClose === -1) return null
+
+    if (nextOpen !== -1 && nextOpen < nextClose) {
+      depth++
+      i = nextOpen + 4
+    } else {
+      depth--
+      i = nextClose + 6
+    }
+  }
+
+  return html.slice(open.index + open[0].length, i - 6)
+}
+
 export function parseHitCount(html: string): number {
   const match = html.match(/af <strong>([\d.]+)<\/strong>/)
   if (!match) return 0

.agents/skills/jobindex-search/cli/tests/parsing.test.ts

@@ -1,5 +1,5 @@
 import { describe, test, expect } from "bun:test";
-import { parseJobCards } from "../src/helpers";
+import { parseJobCards, extractDivContent } from "../src/helpers";
 
 // Minimal jobad-wrapper markup: parseJobCards splits on `jobad-wrapper-<id>`
 // and reads the title from the <h4><a href>…</a></h4> and the company from the
@@ -44,3 +44,55 @@ describe("decodeHtmlEntities (via parseJobCards)", () => {
     expect(c.company).toBe("Nørrebro ApS");
   });
 });
+
+describe("extractDivContent", () => {
+  test("extracts content from simple div", () => {
+    const html = '<div class="job-text">Simple text</div>';
+    expect(extractDivContent(html, "job-text")).toBe("Simple text");
+  });
+
+  test("extracts content with nested divs — the regression case", () => {
+    const html = `<div class="job-text">
+      <div class="highlight">First section</div>
+      <div class="details">Second section with important info</div>
+    </div>`;
+    expect(extractDivContent(html, "job-text")).toBe(
+      '\n      <div class="highlight">First section</div>\n      <div class="details">Second section with important info</div>\n    ',
+    );
+  });
+
+  test("returns null when class not found", () => {
+    expect(extractDivContent("<div>no class</div>", "nonexistent")).toBeNull();
+  });
+
+  test("works with extra attributes on the div", () => {
+    const html = '<div id="desc" class="job-text" data-x="1">Content</div>';
+    expect(extractDivContent(html, "job-text")).toBe("Content");
+  });
+
+  test("handles deeply nested divs (3 levels)", () => {
+    const html = `<div class="job-text">
+      <div>
+        <div>Deep content</div>
+      </div>
+    </div>`;
+    expect(extractDivContent(html, "job-text")).toBe(
+      '\n      <div>\n        <div>Deep content</div>\n      </div>\n    ',
+    );
+  });
+
+  test("handles empty content", () => {
+    const html = '<div class="job-text"></div>';
+    expect(extractDivContent(html, "job-text")).toBe("");
+  });
+
+  test("handles br and other non-div tags", () => {
+    const html = '<div class="job-text">Line1<br>Line2<br>Line3</div>';
+    expect(extractDivContent(html, "job-text")).toBe("Line1<br>Line2<br>Line3");
+  });
+
+  test("escapes special regex characters in class name", () => {
+    const html = '<div class="job-text (special)">Content</div>';
+    expect(extractDivContent(html, "job-text (special)")).toBe("Content");
+  });
+});

.agents/skills/linkedin-search/cli/src/helpers.ts

@@ -68,6 +68,37 @@ export interface JobDetail extends JobCard {
   applyUrl: string | null
 }
 
+/**
+ * Extract the inner HTML of a <div> identified by a CSS class name, correctly
+ * handling nested <div> elements by tracking tag depth.
+ */
+export function extractDivContent(html: string, className: string): string | null {
+  const escaped = className.replace(/[.*+?^${}()|[\]\\]/g, '\\$&')
+  const openRe = new RegExp(`<div[^>]*class="[^"]*${escaped}[^"]*"[^>]*>`, 'i')
+  const open = openRe.exec(html)
+  if (!open) return null
+
+  let i = open.index + open[0].length
+  let depth = 1
+
+  while (depth > 0 && i < html.length) {
+    const nextOpen = html.indexOf('<div', i)
+    const nextClose = html.indexOf('</div>', i)
+
+    if (nextClose === -1) return null
+
+    if (nextOpen !== -1 && nextOpen < nextClose) {
+      depth++
+      i = nextOpen + 4
+    } else {
+      depth--
+      i = nextClose + 6
+    }
+  }
+
+  return html.slice(open.index + open[0].length, i - 6)
+}
+
 /**
  * Convert a Unicode code point to a string. Uses `fromCodePoint` (not
  * `fromCharCode`) so supplementary-plane code points (e.g. emoji, U+1F600)
@@ -180,11 +211,11 @@ export function parseJobDetail(html: string, id: string): JobDetail {
 
   // Rich description block. Keep paragraph/line breaks as newlines.
   let description: string | null = null
-  const desc = html.match(
-    /class="(?:show-more-less-html__markup|description__text[^"]*)"[^>]*>([\s\S]*?)<\/div>/i,
-  )
-  if (desc) {
-    const withBreaks = desc[1]
+  const descHtml =
+    extractDivContent(html, "show-more-less-html__markup") ??
+    extractDivContent(html, "description__text")
+  if (descHtml) {
+    const withBreaks = descHtml
       .replace(/<\s*br\s*\/?>/gi, "\n")
       .replace(/<\/(p|li|ul|ol|div|h\d)>/gi, "\n")
     description = decodeHtmlEntities(stripTags(withBreaks)).replace(/\n{3,}/g, "\n\n").trim() || null

.agents/skills/linkedin-search/cli/tests/parsing.test.ts

@@ -1,5 +1,5 @@
 import { describe, test, expect } from "bun:test";
-import { parseJobCards, parseJobDetail } from "../src/helpers";
+import { parseJobCards, parseJobDetail, extractDivContent } from "../src/helpers";
 
 // Minimal search-card markup: parseJobCards splits on the job-posting URN and
 // needs an id, a base-search-card__title, and a full-link. Everything else is
@@ -53,3 +53,61 @@ describe("decodeHtmlEntities (via parseJobDetail)", () => {
     expect(job.title).toBe("Señor Engineer");
   });
 });
+
+describe("extractDivContent", () => {
+  test("extracts content from simple div", () => {
+    const html = '<div class="description__text">Simple text</div>';
+    expect(extractDivContent(html, "description__text")).toBe("Simple text");
+  });
+
+  test("extracts content with nested divs — the regression case", () => {
+    const html = `<div class="description__text">
+      <div>Requirements:</div>
+      <ul><li>Skill A</li></ul>
+      <div>About Us:</div>
+      <p>We are...</p>
+    </div>`;
+    expect(extractDivContent(html, "description__text")).toBe(
+      '\n      <div>Requirements:</div>\n      <ul><li>Skill A</li></ul>\n      <div>About Us:</div>\n      <p>We are...</p>\n    ',
+    );
+  });
+
+  test("returns null when class not found", () => {
+    expect(extractDivContent("<div>no class</div>", "nonexistent")).toBeNull();
+  });
+
+  test("works with show-more-less-html__markup class", () => {
+    const html = '<div class="show-more-less-html__markup">LinkedIn content</div>';
+    expect(extractDivContent(html, "show-more-less-html__markup")).toBe("LinkedIn content");
+  });
+
+  test("handles deeply nested divs (3 levels)", () => {
+    const html = `<div class="description__text">
+      <div>
+        <div>Deep content</div>
+      </div>
+    </div>`;
+    expect(extractDivContent(html, "description__text")).toBe(
+      '\n      <div>\n        <div>Deep content</div>\n      </div>\n    ',
+    );
+  });
+
+  test("handles empty content", () => {
+    const html = '<div class="description__text"></div>';
+    expect(extractDivContent(html, "description__text")).toBe("");
+  });
+
+  test("parseJobDetail uses extractDivContent and preserves full description", () => {
+    const html = `<div class="description__text">
+      <div>Requirements:</div>
+      <ul><li>5 years Python</li></ul>
+      <div>About Us:</div>
+      <p>We are hiring!</p>
+    </div>`;
+    const job = parseJobDetail(html, "999");
+    expect(job.description).toContain("Requirements:");
+    expect(job.description).toContain("5 years Python");
+    expect(job.description).toContain("About Us:");
+    expect(job.description).toContain("We are hiring!");
+  });
+});