Skip git-ignored files in the length gate, and vendored skill bundles

Three exemptions the first cut missed, all of them files no commit could
ever fix:

- Git-ignored files (todo's exported BACKLOG.md, logs, scratch dumps).
  These are rewritten wholesale by whatever produces them, so splitting one
  is meaningless, and pre-commit never sees them anyway. Applied to
  explicitly-named paths as well as --all, so the two modes cannot disagree
  about the same file.
- Vendored agent-skill bundles (.agents/skills, .claude/skills). A SKILL.md
  is a fixed-name file with frontmatter; splitting one breaks the format.
  Matched by directory name wherever it appears, since these get dropped
  into any repo.
- .utils, where CI checks out this repo. Without it a repo's lint job fails
  on utils' own long files, which that repo cannot fix.

Baseline accordingly drops from 657 to 597. Re-verified all eight fixtures:
over-cap fails, under-cap passes, generated/vendored/wordlist/git-ignored
exempt, prose still capped, and the gate still passes on its own source.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01C7yZz45ezmzCB89Cv5gAtZ

file_length/check.py

@@ -11,6 +11,7 @@ from __future__ import annotations
 
 import argparse
 import os
+import subprocess
 import sys
 from pathlib import Path
 
@@ -61,14 +62,45 @@ def exempt_reason(path: Path, lines: int) -> str | None:
     return None
 
 
+def git_ignored(paths: list[Path], root: Path) -> set[Path]:
+    """The subset of `paths` that git ignores, empty if root is not a repo.
+
+    Git-ignored files are build output or generated data the repo does not
+    own -- an app's own export (todo's BACKLOG.md), a log, a scratch dump.
+    They are rewritten wholesale by whatever produces them, so splitting one
+    is meaningless: the next run restores the original. Pre-commit never sees
+    them either (they cannot be staged), so skipping them makes --all agree
+    with the hook instead of failing on files no commit could ever fix.
+    """
+    if not paths:
+        return set()
+    try:
+        result = subprocess.run(
+            ["git", "-C", str(root), "check-ignore", "--stdin"],
+            input="\n".join(str(p) for p in paths),
+            capture_output=True,
+            text=True,
+            check=False,
+        )
+    except OSError:  # git not installed
+        return set()
+    # 0 = some ignored, 1 = none ignored, 128 = not a repo.
+    if result.returncode not in (0, 1):
+        return set()
+    return {Path(line) for line in result.stdout.splitlines() if line}
+
+
 def iter_all(root: Path):
-    """Every file under root, skipping excluded directories."""
+    """Every file under root, skipping excluded and git-ignored files."""
+    found: list[Path] = []
     for dirpath, dirnames, filenames in os.walk(root, followlinks=False):
         dirnames[:] = [d for d in dirnames if d not in EXCLUDED_DIRS]
         for name in filenames:
             candidate = Path(dirpath) / name
             if not candidate.is_symlink():
-                yield candidate
+                found.append(candidate)
+    ignored = git_ignored(found, root)
+    return [path for path in found if path not in ignored]
 
 
 def main() -> int:
@@ -89,7 +121,13 @@ def main() -> int:
     if args.all:
         targets = list(iter_all(Path.cwd()))
     elif args.paths:
-        targets = args.paths
+        # Explicitly-named paths get the same git-ignore skip as --all, so
+        # naming a file directly cannot report a violation the hook never
+        # would. Without this the two modes disagree on exactly the files
+        # no commit could ever fix (todo's exported BACKLOG.md, logs).
+        existing = [p for p in args.paths if p.is_file()]
+        ignored = git_ignored(existing, Path.cwd())
+        targets = [p for p in existing if p not in ignored]
     else:
         parser.error("give file paths or --all")
 
@@ -118,9 +156,7 @@ def main() -> int:
     for path, lines in sorted(violations, key=lambda item: -item[1]):
         over = lines - MAX_LINES
         print(f"  {path}: {lines} lines (over by {over})", file=sys.stderr)
-    print(
-        "\nSplit them, or see refactor_claude_todo.md in this repo.", file=sys.stderr
-    )
+    print("\nSplit them, or see refactor_claude_todo.md in this repo.", file=sys.stderr)
     return 1
 
 

file_length/config.py

@@ -121,6 +121,10 @@ EXCLUDED_DIRS = frozenset(
         ".ci-mirror-venv",
         ".hippo",
         ".build-deps",
+        # Where CI checks out the shared gate itself (kuhyx/utils). Without
+        # this, a repo's lint job fails on utils' own long files -- a gate
+        # that fails for reasons the repo being gated cannot fix.
+        ".utils",
     }
 )
 
@@ -142,8 +146,18 @@ VENDORED_SUBPATHS = {
         "plugins",
         "ide",
     ),
+    # Agent skill bundles vendored from Anthropic (todo: commit f91311f).
+    # A SKILL.md is a fixed-name file with frontmatter; detail belongs in a
+    # sibling references/ dir, so splitting one into SKILL.part2.md would
+    # break the format rather than improve readability.
+    "todo": (".agents/skills",),
 }
 
+#: Directory names that mark a vendored agent-skill tree wherever they appear.
+#: Keyed by name rather than repo because these bundles get dropped into any
+#: repo that wants them, not just the one they were first seen in.
+VENDORED_ANYWHERE = ("/.agents/skills/", "/.claude/skills/")
+
 #: Whole repos under ~ that are third-party clones, not kuhy's work.
 #: Verified 2026-08-14 by `git remote get-url origin` + commit authors.
 THIRD_PARTY_REPOS = frozenset(
@@ -195,7 +209,13 @@ def is_vendored(path: Path) -> bool:
     parts = path.parts
     if any(part in EXCLUDED_DIRS for part in parts):
         return True
+    # Pre-commit passes paths relative to the repo root ('.agents/skills/x'),
+    # while --all walks absolute ones ('/home/kuhy/todo/.agents/skills/x').
+    # Compare against a leading-slash form so one rule covers both modes.
     posix = path.as_posix()
+    anchored = posix if posix.startswith("/") else f"/{posix}"
+    if any(marker in anchored for marker in VENDORED_ANYWHERE):
+        return True
     for repo, subs in VENDORED_SUBPATHS.items():
         for sub in subs:
             if f"/{repo}/{sub}/" in posix or posix.startswith(f"{repo}/{sub}/"):