fix: fully clean project adoption artifacts

packages/casan-devkit/casan-init.py

@@ -27,6 +27,7 @@ from __future__ import annotations
 
 import argparse
 from contextlib import contextmanager
+import hashlib
 import json
 import os
 import re
@@ -359,6 +360,33 @@ def _load_json_or(path, default):
         return default
 
 
+def _safe_project_path(target, relative):
+    """Resolve a manifest-owned relative path without allowing target escape."""
+    normalized = str(relative).replace("\\", "/").strip("/")
+    if not normalized or os.path.isabs(str(relative)):
+        return None
+    candidate = os.path.abspath(os.path.join(target, *normalized.split("/")))
+    try:
+        if os.path.commonpath((target, candidate)) != target:
+            return None
+    except ValueError:
+        return None
+    return candidate
+
+
+def _file_sha256(path):
+    if not os.path.isfile(path) or os.path.islink(path):
+        return None
+    digest = hashlib.sha256()
+    try:
+        with open(path, "rb") as fh:
+            for chunk in iter(lambda: fh.read(1024 * 1024), b""):
+                digest.update(chunk)
+    except OSError:
+        return None
+    return digest.hexdigest()
+
+
 def _backup_once(path, backups):
     if backups is None:
         return
@@ -814,14 +842,35 @@ def cmd_init(args):
         for path in previous_manifest.get("owned_files", [])
         if isinstance(path, str)
     }
+    previous_owned_hashes = {
+        str(path).replace("\\", "/"): digest
+        for path, digest in previous_manifest.get(
+            "owned_file_hashes", {}).items()
+        if isinstance(path, str) and isinstance(digest, str)
+    }
+    newly_owned_files = set()
+    level_removed = []
+    level_retained = []
+    if not apply_devkit:
+        devkit_owned_files = {
+            path for path in owned_files
+            if (path.startswith(".gitea/workflows/casan-") or
+                path.startswith("apps/"))
+        }
+        level_removed, level_retained = _remove_owned_project_artifacts(
+            target, devkit_owned_files, previous_owned_hashes)
+        for relative in level_removed:
+            owned_files.discard(relative)
+            previous_owned_hashes.pop(relative, None)
 
     def created_add(p):
         created.append(os.path.relpath(p, target))
 
     def mark_owned_if_absent(p):
         if not os.path.exists(p):
-            owned_files.add(
-                os.path.relpath(p, target).replace(os.sep, "/"))
+            relative = os.path.relpath(p, target).replace(os.sep, "/")
+            owned_files.add(relative)
+            newly_owned_files.add(relative)
 
     # ── .casan/config.json ──
     cfg_dir = os.path.join(target, ".casan")
@@ -969,24 +1018,41 @@ def cmd_init(args):
         ci_src = os.path.join(dk, "templates", "gitea-workflow", "ci.yml")
         ci_dst = os.path.join(target, ".gitea", "workflows", "casan-ci.yml")
         if _copy_if_absent(ci_src, ci_dst, None, target, created):
-            owned_files.add(os.path.relpath(
-                ci_dst, target).replace(os.sep, "/"))
+            relative = os.path.relpath(
+                ci_dst, target).replace(os.sep, "/")
+            owned_files.add(relative)
+            newly_owned_files.add(relative)
             level_extras.append(".gitea/workflows/casan-ci.yml")
         dom_src = os.path.join(dk, "templates", "domain-pack")
         dom_dst = os.path.join(target, "apps", project, "domain")
         created_before_domain = len(created)
         n = _copy_tree_missing(dom_src, dom_dst, target, created)
         if n:
-            owned_files.update(
+            domain_files = {
                 path.replace(os.sep, "/")
-                for path in created[created_before_domain:])
+                for path in created[created_before_domain:]
+            }
+            owned_files.update(domain_files)
+            newly_owned_files.update(domain_files)
             level_extras.append("apps/%s/domain (%d files)" % (project, n))
 
     # ── manifest (so uninstall/verify know what init created) ──
+    owned_file_hashes = {
+        path: digest
+        for path, digest in previous_owned_hashes.items()
+        if path in owned_files
+    }
+    for relative in newly_owned_files:
+        owned_path = _safe_project_path(target, relative)
+        digest = _file_sha256(owned_path) if owned_path else None
+        if digest:
+            owned_file_hashes[relative] = digest
     manifest = {
+        "manifest_version": 2,
         "created": created,
         "backups": [os.path.relpath(b, target) for b in backups],
         "owned_files": sorted(owned_files),
+        "owned_file_hashes": dict(sorted(owned_file_hashes.items())),
         "project_id": project,
         "at": now_iso(),
     }
@@ -1008,6 +1074,8 @@ def cmd_init(args):
         "vscode_extension": vscode_install,
         "legacy_migration": legacy_migration,
         "level_extras": level_extras,
+        "level_removed": level_removed,
+        "level_retained": level_retained,
         "note": ("harness NOT copied into repo (hybrid model); selected client hooks "
                  "MERGED and unselected CASAN hooks removed; run `casan doctor`"),
     }
@@ -1300,6 +1368,54 @@ def _prune_empty_directory(path):
         pass
 
 
+def _prune_empty_parents(path, target):
+    parent = os.path.dirname(path)
+    while parent and parent != target:
+        try:
+            os.rmdir(parent)
+        except OSError:
+            break
+        parent = os.path.dirname(parent)
+
+
+def _remove_owned_project_artifacts(target, owned_files, owned_file_hashes):
+    """Remove safe standalone artifacts while preserving merged/user content."""
+    removed = []
+    retained = []
+    integration_owned_files = {
+        ".claude/settings.json",
+        ".codex/hooks.json",
+        ".codex/config.toml",
+        ".vscode/extensions.json",
+        ".specify/.gitignore",
+    }
+    for relative in sorted(
+            owned_files, key=lambda item: (item.count("/"), item),
+            reverse=True):
+        if (relative in integration_owned_files or
+                relative.startswith(".casan/")):
+            continue
+        path = _safe_project_path(target, relative)
+        if not path:
+            retained.append("%s (invalid ownership path)" % relative)
+            continue
+        if not os.path.isfile(path) and not os.path.islink(path):
+            continue
+        expected_hash = owned_file_hashes.get(relative)
+        current_hash = _file_sha256(path)
+        casan_namespaced_workflow = (
+            relative.startswith(".gitea/workflows/casan-") and
+            relative.endswith((".yml", ".yaml")))
+        if casan_namespaced_workflow or (
+                expected_hash and current_hash == expected_hash):
+            _remove_owned_file(path, removed, target)
+            _prune_empty_parents(path, target)
+        else:
+            retained.append(
+                "%s (modified or legacy file; preserved)" % relative)
+    return removed, retained
+
+
 def _uninstall_vscode_extension(enabled):
     if not enabled:
         return {
@@ -1349,6 +1465,11 @@ def cmd_uninstall(args):
         for path in manifest.get("owned_files", [])
         if isinstance(path, str)
     }
+    owned_file_hashes = {
+        str(path).replace("\\", "/"): digest
+        for path, digest in manifest.get("owned_file_hashes", {}).items()
+        if isinstance(path, str) and isinstance(digest, str)
+    }
     hook_changes = {}
     integration_files = (
         (".claude/settings.json", "claude_hook.py", "casan-hook.py"),
@@ -1379,7 +1500,9 @@ def cmd_uninstall(args):
         vscode_path, "_casan_unused", "",
         ".vscode/extensions.json" in owned_files)
 
-    removed = []
+    removed, retained = _remove_owned_project_artifacts(
+        target, owned_files, owned_file_hashes)
+
     for relative in (
             ".casan/config.json",
             ".casan/version.lock",
@@ -1403,14 +1526,8 @@ def cmd_uninstall(args):
             ".casan", ".claude", ".codex", ".vscode", ".specify"):
         _prune_empty_directory(os.path.join(target, relative))
 
-    retained = []
     if not args.purge and os.path.isdir(os.path.join(target, ".specify")):
         retained.append(".specify/ runtime evidence (use --purge to remove)")
-    for item in manifest.get("created", []):
-        normalized = str(item).replace("\\", "/")
-        if (normalized.startswith(".gitea/") or
-                normalized.startswith("apps/")):
-            retained.append(normalized + " (may contain project changes)")
     backups = manifest.get("backups", [])
     if backups:
         retained.append("%d .casan-bak backup(s)" % len(backups))
@@ -1448,8 +1565,8 @@ def main(argv=None):
         "--vscode-install", choices=["auto", "yes", "no"], default="auto",
         help=("install the local CASAN @casan VSIX when vscode-copilot is selected "
               "(default auto: install when `code` is available)"))
-    pi.add_argument("--level", default="devkit",
-                    help="packaging level to adopt: 1|core, 2|devkit (default), 3|platform (preview), 4|enterprise (refused)")
+    pi.add_argument("--level", default="core",
+                    help="packaging level to adopt: 1|core (default), 2|devkit, 3|platform (preview), 4|enterprise (refused)")
     pi.add_argument("--mode", choices=["observe", "enforce"], default="enforce",
                     help="agentic policy mode (default: enforce; use observe for a telemetry-only pilot)")
     pi.add_argument("--integration-mode", dest="integration_mode",