Merge PR #606: fix(compress): never classify build files or shebang scripts as prose (fixes #600)

plugins/caveman/skills/caveman-compress/scripts/detect.py

@@ -17,6 +17,16 @@ SKIP_EXTENSIONS = {
     ".dockerfile", ".makefile", ".csv", ".ini", ".cfg",
 }
 
+# Well-known build/config files that carry no (or a misleading) extension —
+# `Dockerfile` has no suffix so `.dockerfile` above never matches it, and
+# `CMakeLists.txt` would ride the compressible `.txt` rule. Checked by
+# basename before any extension rule.
+KNOWN_CODE_FILENAMES = {
+    "dockerfile", "makefile", "gnumakefile", "jenkinsfile", "vagrantfile",
+    "rakefile", "gemfile", "justfile", "procfile", "brewfile",
+    "cmakelists.txt",
+}
+
 # Patterns that indicate a line is code
 CODE_PATTERNS = [
     re.compile(r"^\s*(import |from .+ import |require\(|const |let |var )"),
@@ -67,6 +77,10 @@ def detect_file_type(filepath: Path) -> str:
     """
     ext = filepath.suffix.lower()
 
+    # Known code filenames win over any extension rule
+    if filepath.name.lower() in KNOWN_CODE_FILENAMES:
+        return "code"
+
     # Extension-based classification
     if ext in COMPRESSIBLE_EXTENSIONS:
         return "natural_language"
@@ -82,6 +96,10 @@ def detect_file_type(filepath: Path) -> str:
 
         lines = text.splitlines()[:50]
 
+        # Shebang means executable script, never prose
+        if text.startswith("#!"):
+            return "code"
+
         if _is_json_content(text[:10000]):
             return "config"
         if _is_yaml_content(lines):

skills/caveman-compress/scripts/detect.py

@@ -17,6 +17,16 @@ SKIP_EXTENSIONS = {
     ".dockerfile", ".makefile", ".csv", ".ini", ".cfg",
 }
 
+# Well-known build/config files that carry no (or a misleading) extension —
+# `Dockerfile` has no suffix so `.dockerfile` above never matches it, and
+# `CMakeLists.txt` would ride the compressible `.txt` rule. Checked by
+# basename before any extension rule.
+KNOWN_CODE_FILENAMES = {
+    "dockerfile", "makefile", "gnumakefile", "jenkinsfile", "vagrantfile",
+    "rakefile", "gemfile", "justfile", "procfile", "brewfile",
+    "cmakelists.txt",
+}
+
 # Patterns that indicate a line is code
 CODE_PATTERNS = [
     re.compile(r"^\s*(import |from .+ import |require\(|const |let |var )"),
@@ -67,6 +77,10 @@ def detect_file_type(filepath: Path) -> str:
     """
     ext = filepath.suffix.lower()
 
+    # Known code filenames win over any extension rule
+    if filepath.name.lower() in KNOWN_CODE_FILENAMES:
+        return "code"
+
     # Extension-based classification
     if ext in COMPRESSIBLE_EXTENSIONS:
         return "natural_language"
@@ -82,6 +96,10 @@ def detect_file_type(filepath: Path) -> str:
 
         lines = text.splitlines()[:50]
 
+        # Shebang means executable script, never prose
+        if text.startswith("#!"):
+            return "code"
+
         if _is_json_content(text[:10000]):
             return "config"
         if _is_yaml_content(lines):

tests/test_detect.py

@@ -0,0 +1,100 @@
+"""Tests for detect.py file-type classification (issue #600).
+
+Extensionless build files (`Dockerfile`, `Makefile`) and shebang scripts
+used to fall through to the content heuristic and come back as
+compressible natural language — so `/caveman-compress Dockerfile` would
+overwrite a Dockerfile with caveman prose. These tests pin the basename
+and shebang guards.
+"""
+
+import sys
+import tempfile
+import unittest
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(REPO_ROOT / "skills" / "caveman-compress"))
+
+from scripts.detect import detect_file_type, should_compress  # noqa: E402
+
+
+DOCKERFILE_BODY = """FROM node:20-bookworm
+WORKDIR /app
+COPY package.json .
+RUN npm install
+CMD ["node", "index.js"]
+"""
+
+MAKEFILE_BODY = """all: build
+
+build:
+\tgo build -o bin/app ./cmd/app
+
+clean:
+\trm -rf bin
+"""
+
+SHEBANG_BODY = """#!/usr/bin/env bash
+set -euo pipefail
+echo "deploying"
+"""
+
+PROSE_BODY = """This project collects notes about our deployment process.
+
+The main goal is to keep the steps simple enough that anyone on the
+team can run a release without asking for help. Start by reading the
+overview, then follow the checklist in order.
+"""
+
+
+class DetectFileTypeTests(unittest.TestCase):
+    def _write(self, dirpath, name, body):
+        p = Path(dirpath) / name
+        p.write_text(body, encoding="utf-8")
+        return p
+
+    def test_dockerfile_is_code(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "Dockerfile", DOCKERFILE_BODY)
+            self.assertEqual(detect_file_type(p), "code")
+            self.assertFalse(should_compress(p))
+
+    def test_makefile_is_code(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "Makefile", MAKEFILE_BODY)
+            self.assertEqual(detect_file_type(p), "code")
+            self.assertFalse(should_compress(p))
+
+    def test_known_names_case_insensitive(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            for name in ("dockerfile", "MAKEFILE", "Jenkinsfile", "Vagrantfile"):
+                p = self._write(tmp, name, "irrelevant body\n")
+                self.assertEqual(detect_file_type(p), "code", name)
+
+    def test_cmakelists_txt_not_compressible_despite_txt_extension(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "CMakeLists.txt", "add_executable(app main.c)\n")
+            self.assertEqual(detect_file_type(p), "code")
+            self.assertFalse(should_compress(p))
+
+    def test_shebang_script_is_code(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "deploy", SHEBANG_BODY)
+            self.assertEqual(detect_file_type(p), "code")
+            self.assertFalse(should_compress(p))
+
+    def test_extensionless_prose_still_compressible(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "NOTES", PROSE_BODY)
+            self.assertEqual(detect_file_type(p), "natural_language")
+            self.assertTrue(should_compress(p))
+
+    def test_markdown_still_compressible(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            p = self._write(tmp, "README.md", PROSE_BODY)
+            self.assertEqual(detect_file_type(p), "natural_language")
+            self.assertTrue(should_compress(p))
+
+
+if __name__ == "__main__":
+    unittest.main()