Fix s07 skill frontmatter parsing edge cases

s07_skill_loading/README.en.md

@@ -70,8 +70,8 @@ def _scan_skills():
         if manifest.exists():
             raw = manifest.read_text()
             meta, body = _parse_frontmatter(raw)
-            name = meta.get("name", d.name)
-            desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+            name = meta.get("name") or d.name
+            desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
             SKILL_REGISTRY[name] = {"name": name, "description": desc, "content": raw}
 
 _scan_skills()  # runs once at startup

s07_skill_loading/README.ja.md

@@ -70,8 +70,8 @@ def _scan_skills():
         if manifest.exists():
             raw = manifest.read_text()
             meta, body = _parse_frontmatter(raw)
-            name = meta.get("name", d.name)
-            desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+            name = meta.get("name") or d.name
+            desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
             SKILL_REGISTRY[name] = {"name": name, "description": desc, "content": raw}
 
 _scan_skills()  # runs once at startup

s07_skill_loading/README.md

@@ -70,8 +70,8 @@ def _scan_skills():
         if manifest.exists():
             raw = manifest.read_text()
             meta, body = _parse_frontmatter(raw)
-            name = meta.get("name", d.name)
-            desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+            name = meta.get("name") or d.name
+            desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
             SKILL_REGISTRY[name] = {"name": name, "description": desc, "content": raw}
 
 _scan_skills()  # runs once at startup

s07_skill_loading/code.py

@@ -52,16 +52,26 @@
 # s07: Skill catalog scan (used by build_system below)
 def _parse_frontmatter(text: str) -> tuple[dict, str]:
     """Parse YAML frontmatter from SKILL.md. Returns (meta, body)."""
-    if not text.startswith("---"):
+    if not (text.startswith("---\n") or text.startswith("---\r\n")):
         return {}, text
-    parts = text.split("---", 2)
-    if len(parts) < 3:
+
+    lines = text.splitlines(keepends=True)
+    closing_index = None
+    for index, line in enumerate(lines[1:], start=1):
+        if line.strip() == "---":
+            closing_index = index
+            break
+    if closing_index is None:
         return {}, text
+
+    frontmatter = "".join(lines[1:closing_index])
+    body = "".join(lines[closing_index + 1 :]).strip()
     try:
-        meta = yaml.safe_load(parts[1]) or {}
+        loaded = yaml.safe_load(frontmatter) or {}
     except yaml.YAMLError:
-        meta = {}
-    return meta, parts[2].strip()
+        loaded = {}
+    meta = loaded if isinstance(loaded, dict) else {}
+    return meta, body
 
 # Build skill registry at startup (used for safe lookup in load_skill)
 SKILL_REGISTRY: dict[str, dict] = {}
@@ -77,8 +87,8 @@ def _scan_skills():
         if manifest.exists():
             raw = manifest.read_text()
             meta, body = _parse_frontmatter(raw)
-            name = meta.get("name", d.name)
-            desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+            name = meta.get("name") or d.name
+            desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
             SKILL_REGISTRY[name] = {"name": name, "description": desc, "content": raw}
 
 _scan_skills()

s08_context_compact/code.py

@@ -29,11 +29,12 @@
 Builds on s07 (skill loading). Usage:
 
     python s08_context_compact/code.py
-    Needs: pip install anthropic python-dotenv + ANTHROPIC_API_KEY in .env
+    Needs: pip install anthropic python-dotenv pyyaml + ANTHROPIC_API_KEY in .env
 """
 
 import ast, json, os, subprocess, time
 from pathlib import Path
+import yaml
 
 try:
     import readline
@@ -57,17 +58,27 @@
 
 # s07: Skill catalog scan (inherited from s07)
 def _parse_frontmatter(text: str) -> tuple[dict, str]:
-    if not text.startswith("---"):
+    """Parse YAML frontmatter from SKILL.md. Returns (meta, body)."""
+    if not (text.startswith("---\n") or text.startswith("---\r\n")):
         return {}, text
-    parts = text.split("---", 2)
-    if len(parts) < 3:
+
+    lines = text.splitlines(keepends=True)
+    closing_index = None
+    for index, line in enumerate(lines[1:], start=1):
+        if line.strip() == "---":
+            closing_index = index
+            break
+    if closing_index is None:
         return {}, text
-    meta = {}
-    for line in parts[1].strip().splitlines():
-        if ":" in line:
-            k, v = line.split(":", 1)
-            meta[k.strip()] = v.strip().strip('"').strip("'")
-    return meta, parts[2].strip()
+
+    frontmatter = "".join(lines[1:closing_index])
+    body = "".join(lines[closing_index + 1 :]).strip()
+    try:
+        loaded = yaml.safe_load(frontmatter) or {}
+    except yaml.YAMLError:
+        loaded = {}
+    meta = loaded if isinstance(loaded, dict) else {}
+    return meta, body
 
 SKILL_REGISTRY: dict[str, dict] = {}
 
@@ -81,8 +92,8 @@ def _scan_skills():
         if manifest.exists():
             raw = manifest.read_text()
             meta, body = _parse_frontmatter(raw)
-            name = meta.get("name", d.name)
-            desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+            name = meta.get("name") or d.name
+            desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
             SKILL_REGISTRY[name] = {"name": name, "description": desc, "content": raw}
 
 _scan_skills()

s20_comprehensive/code.py

@@ -288,16 +288,27 @@ def keep_worktree(name: str) -> str:
 
 
 def _parse_frontmatter(text: str) -> tuple[dict, str]:
-    if not text.startswith("---"):
+    """Parse YAML frontmatter from SKILL.md. Returns (meta, body)."""
+    if not (text.startswith("---\n") or text.startswith("---\r\n")):
         return {}, text
-    parts = text.split("---", 2)
-    if len(parts) < 3:
+
+    lines = text.splitlines(keepends=True)
+    closing_index = None
+    for index, line in enumerate(lines[1:], start=1):
+        if line.strip() == "---":
+            closing_index = index
+            break
+    if closing_index is None:
         return {}, text
+
+    frontmatter = "".join(lines[1:closing_index])
+    body = "".join(lines[closing_index + 1 :]).strip()
     try:
-        meta = yaml.safe_load(parts[1]) or {}
+        loaded = yaml.safe_load(frontmatter) or {}
     except yaml.YAMLError:
-        meta = {}
-    return meta, parts[2].strip()
+        loaded = {}
+    meta = loaded if isinstance(loaded, dict) else {}
+    return meta, body
 
 
 def scan_skills():
@@ -311,9 +322,9 @@ def scan_skills():
         if not manifest.exists():
             continue
         raw = manifest.read_text()
-        meta, _ = _parse_frontmatter(raw)
-        name = meta.get("name", directory.name)
-        desc = meta.get("description", raw.split("\n")[0].lstrip("#").strip())
+        meta, body = _parse_frontmatter(raw)
+        name = meta.get("name") or directory.name
+        desc = meta.get("description") or body.split("\n", 1)[0].lstrip("#").strip()
         SKILL_REGISTRY[name] = {
             "name": name,
             "description": desc,

tests/test_skill_frontmatter_parsing.py

@@ -0,0 +1,106 @@
+import importlib.util
+import os
+import sys
+import tempfile
+import types
+import unittest
+from pathlib import Path
+
+
+REPO_ROOT = Path(__file__).resolve().parents[1]
+SKILL_MODULES = [
+    ("s07", REPO_ROOT / "s07_skill_loading" / "code.py", "_scan_skills"),
+    ("s08", REPO_ROOT / "s08_context_compact" / "code.py", "_scan_skills"),
+    ("s20", REPO_ROOT / "s20_comprehensive" / "code.py", "scan_skills"),
+]
+
+
+def load_skill_module(module_name: str, module_path: Path, temp_cwd: Path):
+    fake_anthropic = types.ModuleType("anthropic")
+
+    class FakeAnthropic:
+        def __init__(self, *args, **kwargs):
+            self.messages = types.SimpleNamespace(create=None)
+
+    fake_dotenv = types.ModuleType("dotenv")
+    setattr(fake_anthropic, "Anthropic", FakeAnthropic)
+    setattr(fake_dotenv, "load_dotenv", lambda override=True: None)
+
+    previous_modules = {
+        "anthropic": sys.modules.get("anthropic"),
+        "dotenv": sys.modules.get("dotenv"),
+    }
+    previous_cwd = Path.cwd()
+    previous_model_id = os.environ.get("MODEL_ID")
+
+    spec = importlib.util.spec_from_file_location(
+        f"{module_name}_frontmatter_test", module_path)
+    if spec is None or spec.loader is None:
+        raise RuntimeError(f"Unable to load {module_path}")
+    module = importlib.util.module_from_spec(spec)
+
+    sys.modules["anthropic"] = fake_anthropic
+    sys.modules["dotenv"] = fake_dotenv
+    try:
+        os.chdir(temp_cwd)
+        os.environ["MODEL_ID"] = "test-model"
+        spec.loader.exec_module(module)
+        return module
+    finally:
+        os.chdir(previous_cwd)
+        if previous_model_id is None:
+            os.environ.pop("MODEL_ID", None)
+        else:
+            os.environ["MODEL_ID"] = previous_model_id
+        for name, previous in previous_modules.items():
+            if previous is None:
+                sys.modules.pop(name, None)
+            else:
+                sys.modules[name] = previous
+
+
+class SkillFrontmatterParsingTests(unittest.TestCase):
+    def test_scan_skills_falls_back_for_empty_metadata_values(self):
+        raw = "---\nname:\ndescription:\n---\n# Body description\n\nDetails"
+        for module_name, module_path, scan_name in SKILL_MODULES:
+            with self.subTest(module=module_name), tempfile.TemporaryDirectory() as tmp:
+                tmp_path = Path(tmp)
+                skill_dir = tmp_path / "skills" / "fallback-skill"
+                skill_dir.mkdir(parents=True)
+                (skill_dir / "SKILL.md").write_text(raw)
+
+                module = load_skill_module(module_name, module_path, tmp_path)
+                module.SKILL_REGISTRY.clear()
+                getattr(module, scan_name)()
+
+                self.assertIn("fallback-skill", module.SKILL_REGISTRY)
+                self.assertEqual(
+                    module.SKILL_REGISTRY["fallback-skill"]["description"],
+                    "Body description",
+                )
+
+    def test_parse_frontmatter_treats_non_mapping_yaml_as_empty_meta(self):
+        raw = "---\n- not\n- a\n- mapping\n---\nBody"
+        for module_name, module_path, _ in SKILL_MODULES:
+            with self.subTest(module=module_name), tempfile.TemporaryDirectory() as tmp:
+                module = load_skill_module(module_name, module_path, Path(tmp))
+
+                meta, body = module._parse_frontmatter(raw)
+
+                self.assertEqual(meta, {})
+                self.assertEqual(body, "Body")
+
+    def test_parse_frontmatter_requires_opening_delimiter_on_own_line(self):
+        raw = "---not frontmatter\n---\n# Body"
+        for module_name, module_path, _ in SKILL_MODULES:
+            with self.subTest(module=module_name), tempfile.TemporaryDirectory() as tmp:
+                module = load_skill_module(module_name, module_path, Path(tmp))
+
+                meta, body = module._parse_frontmatter(raw)
+
+                self.assertEqual(meta, {})
+                self.assertEqual(body, raw)
+
+
+if __name__ == "__main__":
+    unittest.main()