feat: add top starred repos to organization profile

README.md

@@ -11,6 +11,9 @@ change its layout or introduction.
 - Ten most recently published GitHub Releases across all public repos. Drafts
   are excluded; prereleases are labeled. Git tags without a release do not appear.
 - Five latest posts from [the writing RSS feed](https://2389.ai/research/writing/index.xml).
+- Ten public repos with the most stars, with counts, descriptions, and fork/archive
+  labels. Ties sort alphabetically, ignoring case first. All public repos qualify;
+  star counts come from the existing repository requests.
 - An alphabetical, expandable directory of **every public repo**, including
   forks and archived projects. Repository and release requests fetch all pages.
 

docs/superpowers/plans/2026-10-04-profile.md

@@ -41,3 +41,28 @@ Owner: orchestrator after generator interfaces are settled.
 - Independent review: spec, code quality, and workflow integration approved with no findings. Parent fresh-eyes review found no additional issues. Generated Markdown renders through GitHub's Markdown API.
 - Second live generation returned `Unchanged`. GitHub browser inspection confirmed desktop/mobile rendering, an expandable directory with 117 entries, and no horizontal overflow at a 390px viewport.
 - Implementation committed as `da4382b`; [draft PR #1](https://github.com/2389-research/.github/pull/1) opened. Both initial GitHub CI runs passed. Next step: review and merge; the default-branch schedule is not active until then. Compaction count: 0.
+
+## Follow-up: top starred repos
+
+Doctor Biz merged PR #1 and requested a section near the bottom for top-starred repos.
+Add `## Top starred repos` immediately before the repository directory. Show ten
+public repositories sorted by descending `stargazers_count`, then case-insensitive
+name and exact name for stable ties. Include star counts and existing descriptions,
+with the same archived/fork labels as the directory. Use all public repositories;
+do not add filtering beyond the current privacy filter. Reuse the current API
+responses and daily workflow. Validate star counts as nonnegative integers.
+
+- [x] TDD: ranking, numeric counts, tie order, ten-item limit, public-only selection,
+  section placement, invalid counts, and HTTP collection through rendering.
+- [x] Implement the small model/parser/render change; document the section and
+  extend the real API E2E check for star counts and descending order.
+- [x] Run `scripts/check`, live E2E, generate the profile, and inspect its rendering.
+- [x] Review the follow-up for delivery on `feat/top-starred-repos`; leave merging
+  to Doctor Biz.
+
+Validation: 18 offline tests, Ruff format/lint, strict mypy, and actionlint pass.
+Live E2E passes against real APIs; GitHub's Markdown renderer shows ten ranked
+entries immediately before the directory. Review caught and fixed a test that
+compared star counts across separate live reads. Independent re-review and
+fresh-eyes review have no remaining findings. The existing daily job refreshes
+the new section without workflow changes.

gotchas.md

@@ -4,3 +4,4 @@
 - Research posts come from `https://2389.ai/research/writing/index.xml`. The parent research feed was empty when checked on 2026-10-04.
 - “Tagged releases” means published GitHub Releases, confirmed by Doctor Biz. Draft releases and private repositories must never appear.
 - The org already exceeds GitHub's 100-repo page size. Fetch every page before generating the directory.
+- The top-starred section uses `stargazers_count` from the same repository responses. Star counts can change between live API reads; exact ranking expectations belong in stable fixture tests.

profile/README.md

@@ -36,6 +36,19 @@ We're an applied AI lab in Chicago. We build tools for agents, share our researc
 - [Tracker Runner: a control plane for Tracker pipelines](https://2389.ai/research/writing/tracker-runner/) — 2026-09-10
 - [Tracker Grew Up: Two Months of Making Agent Pipelines Boring](https://2389.ai/research/writing/tracker-grew-up/) — 2026-09-08
 
+## Top starred repos
+
+- [claude\-plugins](https://github.com/2389-research/claude-plugins) — 28 plugins and MCP servers for Claude Code — TDD, multi\-agent orchestration, iterative refinement, binary RE, structured decisions\. Install any skill in one command\. — 95 stars
+- [dippin\-lang](https://github.com/2389-research/dippin-lang) — Dippin: a DSL for authoring AI pipeline workflows — 32 stars
+- [landing\-page\-design](https://github.com/2389-research/landing-page-design) — Create high\-converting, visually distinctive landing pages with Vibe Discovery process and anti\-AI\-slop principles — 21 stars
+- [tracker](https://github.com/2389-research/tracker) — 21 stars
+- [binary\-re](https://github.com/2389-research/binary-re) — Agentic binary reverse engineering for ELF binaries on ARM64, ARMv7, x86\_64 \- hypothesis\-driven analysis with radare2, Ghidra, GDB, QEMU — 18 stars
+- [thrifty](https://github.com/2389-research/thrifty) — Tiered\-delegation task execution for Claude Code: strong model plans, cheap models execute in parallel, mid model verifies against acceptance criteria\. — 17 stars
+- [simmer](https://github.com/2389-research/simmer) — Iterative artifact refinement with investigation\-first judge board \- constructs problem\-specific judges that read the code, understand the problem, and propose evidence\-based improvements — 15 stars
+- [mcp\-socialmedia](https://github.com/2389-research/mcp-socialmedia) — 14 stars
+- [summarize\-meetings](https://github.com/2389-research/summarize-meetings) — Batch\-process meeting transcripts from Obsidian vault into structured summaries with knowledge graph updates, people notes, and concept extraction — 13 stars
+- [review\-squad](https://github.com/2389-research/review-squad) — Dispatch panels of specialized subagents to review projects — expert audits, first\-impression personas, task\-completion flows, and pedantic nitpicks — 12 stars
+
 ## Repository directory
 
 <details>

profile/generate.py

@@ -31,6 +31,7 @@ class Repository:
     archived: bool
     fork: bool
     pushed: datetime | None
+    stars: int
 
 
 @dataclass(frozen=True)
@@ -86,6 +87,9 @@ def parse_repositories(records: list[dict[str, Any]]) -> list[Repository]:
         description = record.get("description")
         if description is not None and not isinstance(description, str):
             raise ValueError("Invalid repository description")
+        stars = record.get("stargazers_count")
+        if type(stars) is not int or stars < 0:
+            raise ValueError("Missing or invalid stargazers_count")
         repositories.append(
             Repository(
                 text_field(record, "name"),
@@ -96,6 +100,7 @@ def parse_repositories(records: list[dict[str, Any]]) -> list[Repository]:
                 timestamp(text_field(record, "pushed_at"))
                 if record.get("pushed_at") is not None
                 else None,
+                stars,
             )
         )
     return repositories
@@ -283,6 +288,15 @@ def render_profile(
         )
     if not posts:
         lines.append("No research posts yet.")
+    lines.extend(["", "## Top starred repos", ""])
+    top_starred = sorted(
+        repositories, key=lambda r: (-r.stars, r.name.casefold(), r.name)
+    )[:10]
+    lines.extend(
+        f"{repository_line(repo)} — {repo.stars} stars" for repo in top_starred
+    )
+    if not top_starred:
+        lines.append("No public repositories yet.")
     lines.extend(
         [
             "",

tests/e2e_generate.py

@@ -20,10 +20,10 @@ def test_public_sources_generate_complete_profile(self) -> None:
         if token:
             headers["Authorization"] = f"Bearer {token}"
         url = "https://api.github.com/orgs/2389-research/repos?type=public&per_page=100"
-        names: list[str] = []
+        names: set[str] = set()
         while url:
             with urlopen(Request(url, headers=headers), timeout=30) as response:
-                names.extend(
+                names.update(
                     repo["name"] for repo in json.load(response) if not repo["private"]
                 )
                 next_page = re.search(
@@ -54,6 +54,24 @@ def test_public_sources_generate_complete_profile(self) -> None:
                 self.assertIn(
                     f"](https://github.com/2389-research/{name})", directory_section
                 )
+            self.assertIn("## Top starred repos", content)
+            top_starred = content.split("## Top starred repos")[1].split("## ")[0]
+            ranked = re.findall(
+                r"^- \[.*\]\(https://github.com/2389-research/([^)]*)\)"
+                r".* — (\d+) stars$",
+                top_starred,
+                re.MULTILINE,
+            )
+            ranking = [(name, int(count)) for name, count in ranked]
+            self.assertEqual(len(ranking), min(10, len(names)))
+            self.assertEqual(len({name for name, _ in ranking}), len(ranking))
+            for name, count in ranking:
+                self.assertIn(name, names)
+                self.assertGreaterEqual(count, 0)
+            self.assertEqual(
+                ranking,
+                sorted(ranking, key=lambda row: (-row[1], row[0].casefold(), row[0])),
+            )
             releases = content.split("## Latest releases")[1].split("## ")[0]
             self.assertIn("/releases/tag/", releases)
             self.assertGreater(releases.count("- ["), 0)

tests/test_generate.py

@@ -28,6 +28,7 @@ def repo(name: str, **fields: Any) -> dict[str, Any]:
         "archived": False,
         "fork": False,
         "pushed_at": "2026-01-01T00:00:00Z",
+        "stargazers_count": 0,
         **fields,
     }
 
@@ -77,6 +78,69 @@ def test_directory_is_complete_sorted_and_labels_archives_and_forks(self) -> Non
         self.assertIn("fork", directory)
         self.assertNotIn("private", directory)
 
+    def test_top_starred_ranks_numerically_limits_to_ten_and_breaks_ties(self) -> None:
+        records = [repo(f"repo-{n}", stargazers_count=n) for n in range(10)]
+        records += [
+            repo("zeta", stargazers_count=100),
+            repo("beta", stargazers_count=20),
+            repo("alpha", stargazers_count=20),
+            repo("Alpha", stargazers_count=20),
+        ]
+        output = generate.render_profile(generate.parse_repositories(records), [], [])
+        self.assertIn("## Top starred repos", output)
+        section = output.split("## Top starred repos")[1].split("## ")[0]
+        lines = [line for line in section.splitlines() if line.startswith("- [")]
+        expected = ["zeta", "Alpha", "alpha", "beta"] + [
+            f"repo-{n}" for n in range(9, 3, -1)
+        ]
+        self.assertEqual(len(lines), 10)
+        for line, name in zip(lines, expected, strict=True):
+            self.assertIn(f"](https://github.com/2389-research/{name})", line)
+        self.assertTrue(lines[0].endswith(" — 100 stars"))
+        self.assertTrue(lines[-1].endswith(" — 4 stars"))
+        self.assertEqual(
+            output.split("## Top starred repos")[1].split("## ")[1].splitlines()[0],
+            "Repository directory",
+        )
+
+    def test_top_starred_includes_all_public_repo_types_and_existing_labels(
+        self,
+    ) -> None:
+        records = [
+            repo("archive", archived=True, stargazers_count=4),
+            repo("fork", fork=True, stargazers_count=3),
+            repo(".github", stargazers_count=2),
+            repo("empty", pushed_at=None, stargazers_count=0),
+            repo("secret", private=True, stargazers_count=100),
+        ]
+        output = generate.render_profile(generate.parse_repositories(records), [], [])
+        self.assertIn("## Top starred repos", output)
+        section = output.split("## Top starred repos")[1].split("## ")[0]
+        self.assertEqual(section.count("- ["), 4)
+        self.assertIn("(archived) — Useful code — 4 stars", section)
+        self.assertIn("(fork) — Useful code — 3 stars", section)
+        self.assertIn("/2389-research/.github)", section)
+        self.assertIn("/2389-research/empty) — Useful code — 0 stars", section)
+        self.assertNotIn("secret", section)
+
+    def test_repository_star_counts_must_be_nonnegative_integers(self) -> None:
+        for count in [-1, 1.5, "5", None, True, False]:
+            with (
+                self.subTest(count=count),
+                self.assertRaisesRegex(ValueError, "stargazers_count"),
+            ):
+                generate.parse_repositories([repo("alpha", stargazers_count=count)])
+        record = repo("alpha")
+        del record["stargazers_count"]
+        with self.assertRaisesRegex(ValueError, "stargazers_count"):
+            generate.parse_repositories([record])
+
+    def test_top_starred_empty_state(self) -> None:
+        output = generate.render_profile([], [], [])
+        self.assertIn("## Top starred repos", output)
+        section = output.split("## Top starred repos")[1].split("## ")[0]
+        self.assertIn("No public repositories yet.", section)
+
     def test_recent_activity_uses_push_time_and_limits_to_eight_active_repos(
         self,
     ) -> None:
@@ -302,6 +366,32 @@ def test_bad_or_failed_sources_preserve_output(self) -> None:
                 )
             self.assertEqual(path.read_text(), "last good profile\n")
 
+    def test_refresh_renders_star_counts_from_every_repository_page(self) -> None:
+        self.respond(
+            "/orgs/2389-research/repos?type=public&per_page=100",
+            [repo("alpha", stargazers_count=9)],
+            Link=f'<{self.base}/page2>; rel="next"',
+        )
+        self.respond(
+            "/page2",
+            [
+                repo("beta", stargazers_count=100, archived=True, fork=True),
+                repo("secret", private=True, stargazers_count=999),
+            ],
+        )
+        self.respond("/repos/2389-research/beta/releases?per_page=100", [])
+        with tempfile.TemporaryDirectory() as directory:
+            path = Path(directory) / "README.md"
+            generate.refresh(path, api_url=self.base, feed_url=f"{self.base}/feed")
+            output = path.read_text()
+        self.assertIn("## Top starred repos", output)
+        section = output.split("## Top starred repos")[1].split("## ")[0]
+        self.assertLess(section.index("beta"), section.index("alpha"))
+        self.assertIn("(archived, fork) — Useful code — 100 stars", section)
+        self.assertIn("Useful code — 9 stars", section)
+        self.assertNotIn("secret", output)
+        self.assertEqual(len(self.requests), 5)
+
     def test_http_failure_reports_status_without_retaining_open_response(self) -> None:
         self.responses["/failure"] = (503, b"unavailable", {})
         with self.assertRaisesRegex(OSError, "HTTP 503 from") as failure: