lintai docs

Appearance

Rule Reference

SEC372lintai-ai-securitystableclaude_settingswarn

Claude settings: wildcard Read permissions

Claude settings permissions allow `Read(*)` in a shared committed config

Provider
lintai-ai-security
Surface
claude_settings
Scope
per_file
Tier
stable
Severity
warn
Confidence
high
Detection
structural
Remediation
message_only

Activation Model

Preset Membership

This rule is part of the builtin activation graph through these preset memberships.

previewmembership
claudemembership

Lifecycle

Stable Lifecycle Contract

State

stable_gated

Graduation rationale

Checks shared Claude settings permissions for explicit wildcard `Read(*)` grants.

Deterministic signal basis

ClaudeSettingsSignals exact string detection for `Read(*)` inside permissions.allow on parsed Claude settings JSON.

Malicious corpus
claude-settings-read-wildcard
Benign corpus
claude-settings-read-specific-safe
structured evidence required remediation reviewed
Canonical note

Structural stable rule intended as a high-precision check with deterministic evidence.

Nearby Signals

Related Rules

SEC362Claude settings: wildcard Bash permissionsClaude settings permissions allow `Bash(*)` in a shared committed configclaude_settingsstable
SEC364Claude settings: bypassPermissions default modeClaude settings set `permissions.defaultMode` to `bypassPermissions` in a shared committed configclaude_settingsstable
SEC367Claude settings: wildcard WebFetch permissionsClaude settings permissions allow `WebFetch(*)` in a shared committed configclaude_settingsstable
SEC369Claude settings: wildcard Write permissionsClaude settings permissions allow `Write(*)` in a shared committed configclaude_settingsstable
SEC373Claude settings: wildcard Edit permissionsClaude settings permissions allow `Edit(*)` in a shared committed configclaude_settingsstable
SEC374Claude settings: wildcard WebSearch permissionsClaude settings permissions allow `WebSearch(*)` in a shared committed configclaude_settingsstable

Why It Matters ​

SEC372 flags committed Claude settings files when permissions.allow contains the exact wildcard read grant Read(*).

This is useful because:

  • shared .claude/settings.json files often become team-wide policy by copy/paste
  • Read(*) grants broad file read access instead of a reviewed allowlist
  • broad read access is much harder to defend in code review than scoped patterns tied to known repo paths

What Triggers ​

This rule applies only to committed Claude settings surfaces:

  • .claude/settings.json
  • claude/settings.json

It triggers when:

  • permissions.allow contains the exact string Read(*)

It does not trigger on:

  • narrower reviewed patterns such as Read(./docs/**)
  • fixture-like test/example paths

Examples ​

Bad:

json
{
  "permissions": {
    "allow": ["Read(*)", "Bash(git status)"]
  }
}

Better:

json
{
  "permissions": {
    "allow": ["Read(./docs/**)", "Bash(git status)"]
  }
}

Remediation ​

  • replace Read(*) with specific reviewed read patterns
  • keep shared read permissions scoped to the smallest set of repo paths the workflow actually needs