fix(scan): exclude Python virtual environments from manifest collection + bump Coana CLI to 15.5.10 (1.1.128)

CHANGELOG.md

@@ -4,6 +4,14 @@ All notable changes to this project will be documented in this file.
 
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/).
 
+## [1.1.128](https://github.com/SocketDev/socket-cli/releases/tag/v1.1.128) - 2026-06-25
+
+### Changed
+- Updated the Coana CLI to v `15.5.10`.
+
+### Fixed
+- Scans now skip Python virtual environments when collecting manifest files. Folders named `.venv`, and any folder containing a `pyvenv.cfg` marker (covering `venv`, `env`, and custom-named environments), are excluded — so `socket scan`, reachability, and `socket fix` stay focused on your project's own manifests instead of the thousands installed inside a virtualenv.
+
 ## [1.1.127](https://github.com/SocketDev/socket-cli/releases/tag/v1.1.127) - 2026-06-24
 
 ### Changed

package.json

@@ -1,6 +1,6 @@
 {
   "name": "socket",
-  "version": "1.1.127",
+  "version": "1.1.128",
   "description": "CLI for Socket.dev",
   "homepage": "https://github.com/SocketDev/socket-cli",
   "license": "MIT",
@@ -96,7 +96,7 @@
     "@babel/preset-typescript": "7.27.1",
     "@babel/runtime": "7.28.4",
     "@biomejs/biome": "2.2.4",
-    "@coana-tech/cli": "15.5.9",
+    "@coana-tech/cli": "15.5.10",
     "@cyclonedx/cdxgen": "12.1.2",
     "@dotenvx/dotenvx": "1.49.0",
     "@eslint/compat": "1.3.2",

src/utils/glob.mts

@@ -36,10 +36,18 @@ export const IGNORED_DIRS = [
   // Taken from globby:
   // https://github.com/sindresorhus/globby/blob/v14.0.2/ignore.js#L11-L16
   'flow-typed',
+  // Conventional Python virtual environment dir. Arbitrarily-named venvs are
+  // detected via their pyvenv.cfg marker during the discovery walk below.
+  '.venv',
 ] as const
 
 const IGNORED_DIR_PATTERNS = IGNORED_DIRS.map(i => `**/${i}`)
 
+// Marker file at the root of every Python virtual environment (stdlib `venv`
+// per PEP 405, and virtualenv >= 20). Lets us detect venvs that don't use a
+// conventional directory name.
+const PYVENV_CFG = 'pyvenv.cfg'
+
 async function getWorkspaceGlobs(
   agent: Agent,
   cwd = process.cwd(),
@@ -251,38 +259,53 @@ export async function globWithGitIgnore(
     ignores.add(pattern)
   }
 
-  // The .gitignore discovery walk has to honor the same directory exclusions
-  // as the package walk below. Otherwise an unreadable subtree (e.g. a
-  // postgres `pgdata` dir owned by another uid, or a Docker volume mount) makes
-  // fast-glob throw `EACCES: permission denied, scandir` *here* — before
-  // --exclude-paths (`cliMinimatchIgnores`) or projectIgnorePaths are ever
-  // applied to the main walk, which is why excluding the path did not help.
-  // `suppressErrors` is the backstop: a directory the user simply cannot read
-  // cannot contain manifests they could scan anyway, so skip it instead of
-  // aborting the whole `socket fix` / `socket scan` run. Negated patterns are
-  // dropped — for a discovery walk they could only re-include a subtree (never
-  // prevent a crash), and fast-glob treats `!` ignore entries inconsistently.
-  const gitIgnoreStream = fastGlob.globStream(['**/.gitignore'], {
-    absolute: true,
-    cwd,
-    dot: true,
-    ignore: [
-      ...DEFAULT_IGNORE_FOR_GIT_IGNORE,
-      ...projectIgnoreGlobs,
-      ...cliMinimatchIgnores,
-    ]
-      .filter(p => p.charCodeAt(0) !== 33 /*'!'*/)
-      .map(stripTrailingSlash),
-    suppressErrors: true,
-  })
+  // The discovery walk (`.gitignore` files plus `pyvenv.cfg` venv markers) has
+  // to honor the same directory exclusions as the package walk below. Otherwise
+  // an unreadable subtree (e.g. a postgres `pgdata` dir owned by another uid, or
+  // a Docker volume mount) makes fast-glob throw `EACCES: permission denied,
+  // scandir` *here* — before --exclude-paths (`cliMinimatchIgnores`) or
+  // projectIgnorePaths are ever applied to the main walk, which is why excluding
+  // the path did not help. `suppressErrors` is the backstop: a directory the
+  // user simply cannot read cannot contain manifests they could scan anyway, so
+  // skip it instead of aborting the whole `socket fix` / `socket scan` run.
+  // Negated patterns are dropped — for a discovery walk they could only
+  // re-include a subtree (never prevent a crash), and fast-glob treats `!`
+  // ignore entries inconsistently. Folding pyvenv.cfg discovery into this same
+  // walk avoids a second full-tree traversal.
+  const discoveryStream = fastGlob.globStream(
+    ['**/.gitignore', `**/${PYVENV_CFG}`],
+    {
+      absolute: true,
+      cwd,
+      dot: true,
+      ignore: [
+        ...DEFAULT_IGNORE_FOR_GIT_IGNORE,
+        ...projectIgnoreGlobs,
+        ...cliMinimatchIgnores,
+      ]
+        .filter(p => p.charCodeAt(0) !== 33 /*'!'*/)
+        .map(stripTrailingSlash),
+      suppressErrors: true,
+    },
+  )
   for await (const ignorePatterns of transform(
-    gitIgnoreStream,
-    async (filepath: string) =>
-      ignoreFileToGlobPatterns(
+    discoveryStream,
+    async (filepath: string) => {
+      if (path.basename(filepath) === PYVENV_CFG) {
+        // A pyvenv.cfg sits at the venv root, so exclude the whole directory.
+        const relDir = path
+          .relative(cwd, path.dirname(filepath))
+          .replace(/\\/g, '/')
+        // An empty relDir means the scan target itself is a venv root; don't
+        // emit `/**`, which would exclude everything the user explicitly targeted.
+        return relDir ? [`${relDir}/**`] : []
+      }
+      return ignoreFileToGlobPatterns(
         (await safeReadFile(filepath)) ?? '',
         filepath,
         cwd,
-      ),
+      )
+    },
     { concurrency: 8 },
   )) {
     for (const p of ignorePatterns) {

src/utils/glob.test.mts

@@ -321,6 +321,79 @@ describe('glob utilities', () => {
         }
       },
     )
+
+    it('excludes a Python virtual environment detected via pyvenv.cfg', async () => {
+      // A venv can use any directory name; the reliable signal is the
+      // pyvenv.cfg marker at its root. Manifests inside it must not surface.
+      mockTestFs({
+        [`${mockFixturePath}/requirements.txt`]: '',
+        [`${mockFixturePath}/myenv/pyvenv.cfg`]: 'home = /usr/bin\nversion = 3.11.0\n',
+        [`${mockFixturePath}/myenv/requirements.txt`]: '',
+        [`${mockFixturePath}/myenv/lib/python3.11/site-packages/foo/setup.py`]:
+          '',
+      })
+
+      const results = await globWithGitIgnore(
+        ['**/requirements.txt', '**/setup.py'],
+        { cwd: mockFixturePath },
+      )
+
+      expect(results.map(normalizePath).sort()).toEqual([
+        `${mockFixturePath}/requirements.txt`,
+      ])
+    })
+
+    it('excludes a `.venv` directory by name', async () => {
+      mockTestFs({
+        [`${mockFixturePath}/package.json`]: '{}',
+        [`${mockFixturePath}/.venv/lib/site-packages/foo/package.json`]: '{}',
+      })
+
+      const results = await globWithGitIgnore(['**/*.json'], {
+        cwd: mockFixturePath,
+      })
+
+      expect(results.map(normalizePath).sort()).toEqual([
+        `${mockFixturePath}/package.json`,
+      ])
+    })
+
+    it('keeps a non-venv directory named `venv` without a pyvenv.cfg', async () => {
+      // Guards against over-exclusion: a bare `venv` dir is only skipped when
+      // it actually contains a pyvenv.cfg, never by name alone.
+      mockTestFs({
+        [`${mockFixturePath}/package.json`]: '{}',
+        [`${mockFixturePath}/venv/package.json`]: '{}',
+      })
+
+      const results = await globWithGitIgnore(['**/*.json'], {
+        cwd: mockFixturePath,
+      })
+
+      expect(results.map(normalizePath).sort()).toEqual([
+        `${mockFixturePath}/package.json`,
+        `${mockFixturePath}/venv/package.json`,
+      ])
+    })
+
+    it('excludes a venv via pyvenv.cfg through the streaming filter path', async () => {
+      // The actual manifest-scan path always passes a filter, so verify the
+      // venv exclusion prunes there too.
+      mockTestFs({
+        [`${mockFixturePath}/package.json`]: '{}',
+        [`${mockFixturePath}/env/pyvenv.cfg`]: 'home = /usr/bin\n',
+        [`${mockFixturePath}/env/lib/site-packages/bar/package.json`]: '{}',
+      })
+
+      const results = await globWithGitIgnore(['**/*'], {
+        cwd: mockFixturePath,
+        filter: filterJsonFiles,
+      })
+
+      expect(results.map(normalizePath).sort()).toEqual([
+        `${mockFixturePath}/package.json`,
+      ])
+    })
   })
 
   describe('createSupportedFilesFilter()', () => {