技能 skill-permissions
📦
skill-permissions
內容修訂 r1 高風險 ⚙️ 外部命令📁 檔案系統存取
稽核技能權限
Claude Code 技能可能需要難以手動審查的 Shell 權限。此技能會分析宣告的命令使用情況，並協助建立相符的允許清單項目，同時針對範圍較廣的變更提出警告。
支援: Claude Codex Code(CC)
安裝前請停止並請求確認。
請審查安裝計畫，並在變更檔案前取得使用者明確同意。
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
智能體請求
Review the Skillstore skill "skill-permissions" from https://skillstore.io/skills/guo-yu-skill-permissions.md and its manifest at https://skillstore.io/api/skills/guo-yu-skill-permissions/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「skill-permissions」。 Analyze the port-allocator skill.
預期結果:
- 列出偵測到的命令，例如 find、ls、cat 和 lsof。
- 顯示建議的 Claude Code allow 規則。
- 建議該技能適用的特定 allow 命令。
正在使用「skill-permissions」。 Show permissions for all skills.
預期結果:
- 顯示已安裝技能與主要權限群組的表格。
- 標記需要授權的技能。
- 建議在批次授權前先審查結果。
正在使用「skill-permissions」。 Authorize one reviewed skill.
預期結果:
- 回報權限規則已合併。
- 顯示受影響的設定檔。
- 提醒使用者權限會在新的工作階段或清除後套用。
安全審計
高風險v6 • 2026/7/9 開啟版本化報告
The skill is a security-sensitive permission management helper, not malware. Most dangerous command hits are documented blocklist examples, but the skill can broaden Claude Code Bash allow rules and batch-modify settings or skill files, so publication should require strong warnings and safer defaults.
3
已掃描檔案
894
分析行數
115
複核項目
0
已忽略的誤報
已確認的安全問題 (4)
高
Batch Permission Expansion and Skill File Modification
The skill documents allow-all and inject modes that can modify settings.json or add allow commands across installed skills. This creates broad local permission expansion risk even though the documentation warns users.
The behavior is explicitly described in multiple language files and affects configuration plus many skill files. The warnings reduce surprise but do not remove the risk.
高
Overbroad Claude Directory Allow Rules
The generated allow rules include broad access to ~/.claude paths, including Bash(cat ~/.claude/*). These rules can expose sensitive local Claude configuration if granted too broadly.
The allow rules are shown directly in the documented authorization flow. The risk depends on user approval, but the scope is broad and security-sensitive.
中
System reconnaissance
- Avoid `inject` or `allow-all` unless you understand the impact
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
System reconnaissance
1. **Conservative analysis** - Only identify explicitly appearing commands, avoid over-authorization
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
能力複核項目 (115)
這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。
高
Hidden file in home directory
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
場所: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
設定ファイル: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **権限設定**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **スキルディレクトリ**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
1. `~/.claude/settings.json` を読み取り
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
SKILL_PATH=$(find ~/.claude/skills -type d -name "<skill-name>" 2>/dev/null | head -1)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
Location: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
Config file: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **Permission config**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **Skill directory**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
1. Read `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
SKILL_PATH=$(find ~/.claude/skills -type d -name "<skill-name>" 2>/dev/null | head -1)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
cat ~/.claude/settings.json 2>/dev/null || echo '{"permissions":{"allow":[]}}'
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
1. 读取 `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
| `cat ~/.claude/*` | `Bash(cat ~/.claude/*)` |
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
📁 位置: ~/.claude/skills/dangerous-skill
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
📁 位置: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
配置文件: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **权限配置**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- **skill 目录**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
1. 读取 `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(find ~/.claude/skills *)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
高
Hidden file in home directory
配置文件: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
場所: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
設定ファイル: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **権限設定**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **スキルディレクトリ**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
1. `~/.claude/settings.json` を読み取り
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
| `/skill-permissions inject` | High | Batch modify all skill SKILL.md files |
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
| `/skill-permissions allow-all` | Medium | Batch modify settings.json |
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
- Avoid `inject` or `allow-all` unless you understand the impact
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
| `/skill-permissions` | Analyze all skills and show permission summary |
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
| `/skill-permissions allow-all` | Execute one-time authorization for all skills |
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
| `/skill-permissions inject` | Inject allow command into all skills |
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
Convert identified commands to `permissions.allow` rules:
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
- **Permission config**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
- **Skill directory**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
1. Read `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
2. Merge the following permissions to `permissions.allow`:
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Shell command substitution
SKILL_PATH=$(find ~/.claude/skills -type d -name "<skill-name>" 2>/dev/null | head -1)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
SKILL_PATH=$(find ~/.claude/skills -type d -name "<skill-name>" 2>/dev/null | head -1)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
Location: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
Config file: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **Permission config**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **Skill directory**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
1. Read `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
### 命令: `/skill-permissions allow-all`
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
### 命令: `/skill-permissions inject`
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
为所有 skill 注入 `allow` 子命令:
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
| `cat ~/.claude/*` | `Bash(cat ~/.claude/*)` |
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
- `allow_override`: 从默认黑名单中移除的命令(谨慎使用)
The documentation allows users to remove patterns from the default blocklist. That is intentional configuration, but it can weaken safeguards around dangerous commands.
中
Ruby/shell backtick execution
- **权限配置**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
- **skill 目录**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
1. **文件修改风险** - `inject` 命令会修改所有 skill 的 SKILL.md 文件
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
2. **权限扩大风险** - `allow-all` 会为所有 skill 添加权限,可能授予过多权限
The documented command can modify many skill files or settings entries. This is intentional functionality, but it creates real permission-expansion and file-modification risk if used without review.
中
Ruby/shell backtick execution
1. 读取 `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Ruby/shell backtick execution
2. 合并以下权限到 `permissions.allow`:
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
cat ~/.claude/* → Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
find ~/.claude/skills -name "SKILL.md" -type f 2>/dev/null
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
SKILL_PATH=$(find ~/.claude/skills -type d -name "<skill-name>" 2>/dev/null | head -1)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
cat ~/.claude/settings.json 2>/dev/null || echo '{"permissions":{"allow":[]}}'
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
1. 读取 `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
| `cat ~/.claude/*` | `Bash(cat ~/.claude/*)` |
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
📁 位置: ~/.claude/skills/dangerous-skill
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
📁 位置: ~/.claude/skills/port-allocator
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
配置文件: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **权限配置**: `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- **skill 目录**: `~/.claude/skills/`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
1. 读取 `~/.claude/settings.json`
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(find ~/.claude/skills *)",
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
"Bash(cat ~/.claude/*)"
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(find ~/.claude/skills *)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
- Bash(cat ~/.claude/*)
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
中
Hidden file access
配置文件: ~/.claude/settings.json
The finding references reading or writing Claude Code settings or skill directories. This is intended permission-management behavior, but it is security-sensitive because it can broaden local tool permissions.
風險因素
⚙️ 外部命令 (50)
SKILL.ja.md:138-139 SKILL.ja.md:107 SKILL.md:16 SKILL.md:17 SKILL.md:18 SKILL.md:19 SKILL.md:22 SKILL.md:23 SKILL.md:24 SKILL.md:30 SKILL.md:31 SKILL.md:32 SKILL.md:33 SKILL.md:34 SKILL.md:35 SKILL.md:44-51 SKILL.md:51-54 SKILL.md:54 SKILL.md:55-60 SKILL.md:60-62 SKILL.md:62-73 SKILL.md:73-77 SKILL.md:77-82 SKILL.md:82-84 SKILL.md:84-94 SKILL.md:94-99 SKILL.md:99-101 SKILL.md:101-107 SKILL.md:107-124 SKILL.md:124-125 SKILL.md:125-126 SKILL.md:126-127 SKILL.md:127-132 SKILL.md:132-133 SKILL.md:133-134 SKILL.md:134-135 SKILL.md:135-140 SKILL.md:140-141 SKILL.md:141-146 SKILL.md:146-162 SKILL.md:162-163 SKILL.md:163-166 SKILL.md:166-180 SKILL.md:180-184 SKILL.md:184-185 SKILL.md:185-190 SKILL.md:190-192 SKILL.md:192-194 SKILL.md:194-198 SKILL.md:198-199
📁 檔案系統存取 (50)
SKILL.ja.md:47 SKILL.ja.md:67 SKILL.ja.md:82 SKILL.ja.md:126 SKILL.ja.md:131 SKILL.ja.md:136 SKILL.ja.md:150 SKILL.ja.md:152 SKILL.ja.md:159 SKILL.ja.md:160 SKILL.ja.md:173 SKILL.ja.md:180 SKILL.ja.md:181 SKILL.ja.md:47 SKILL.ja.md:67 SKILL.ja.md:82 SKILL.ja.md:126 SKILL.ja.md:131 SKILL.ja.md:136 SKILL.ja.md:150 SKILL.ja.md:152 SKILL.ja.md:159 SKILL.ja.md:160 SKILL.ja.md:173 SKILL.ja.md:180 SKILL.ja.md:181 SKILL.ja.md:82 SKILL.md:47 SKILL.md:68 SKILL.md:83 SKILL.md:100 SKILL.md:148 SKILL.md:153 SKILL.md:159 SKILL.md:174 SKILL.md:177 SKILL.md:184 SKILL.md:185 SKILL.md:198 SKILL.md:205 SKILL.md:206 SKILL.md:47 SKILL.md:68 SKILL.md:83 SKILL.md:100 SKILL.md:148 SKILL.md:153 SKILL.md:159 SKILL.md:174 SKILL.md:177
偵測到的模式
System reconnaissance×2
審計單位: claude 檢視審計紀錄 →
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
https://skillstore.io/skills/guo-yu-skill-permissions/audits/6?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/guo-yu-skill-permissions?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/guo-yu-skill-permissions?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/guo-yu-skill-permissions/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/guo-yu-skill-permissions.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>學術引用 (APA · BibTeX · CFF)
APA 引用
guo-yu. (2026). skill-permissions security audit report (audit version 6) [Author version unspecified]. Skillstore. https://skillstore.io/skills/guo-yu-skill-permissions/audits/6BibTeX 引用
@techreport{guo-yu-guo-yu-skill-permissions-2026,
author = {guo-yu},
title = {skill-permissions security audit report (audit version 6)},
institution = {Skillstore},
year = {2026},
number = {6},
url = {https://skillstore.io/skills/guo-yu-skill-permissions/audits/6},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "skill-permissions security audit report (audit version 6)"
version: "unspecified"
type: report
authors:
- name: "guo-yu"
date-released: "2026-07-09"
url: "https://skillstore.io/skills/guo-yu-skill-permissions/audits/6"
identifiers:
- type: other
value: "skillstore:guo-yu-skill-permissions:audit:6"
description: "Skillstore immutable audit report identifier"
Skillstore 評分
評分依據 證據置信度: 中 55
架構
85
可維護性
87
內容
71
社群
83
規範符合度
你可以打造什麼
審查新的社群技能
在授予 Claude Code Bash 權限之前，先分析單一已下載的技能。
準備本機技能集
在團隊設定工作階段之前，彙整已安裝技能的權限摘要。
記錄權限需求
產生 allow 命令指引，說明技能所需的權限。
試試這些提示
分析一項技能
Use skill-permissions to analyze the permissions required by <skill-name> and explain each requested permission.
授權前審查
Run the read-only skill permission analysis for <skill-name>. Do not modify settings until I approve the proposed allow rules.
稽核所有已安裝技能
Analyze all installed Claude Code skills and group the results by low, medium, and high permission risk.
規劃受控授權
Analyze all skills, identify permissions that should not be auto-approved, and prepare explicit allow commands for only the reviewed skills.
最佳實務
- 在使用任何授權命令前，先執行唯讀分析。
- 優先使用明確的逐技能授權，而非 allow-all 或 inject 模式。
- 將產生的萬用字元權限新增到 settings.json 前，請先審查。
避免
- 不要對未經審查的社群技能執行批次授權。
- 不要將產生的 allow 規則視為技能安全的證明。
- 不要在沒有版本控制或備份的情況下，將 allow 命令注入大量技能檔案。
常見問題
此技能會執行偵測到的命令嗎？
文件化的分析流程會掃描技能檔案並產生權限規則。授權命令會修改 Claude Code 設定。
為什麼此技能對安全性很敏感？
它可以新增 Bash allow 規則，也可以在批次模式中修改設定或技能文件。
我應該使用 allow-all 嗎？
只有在審查每一項產生的權限後，才使用 allow-all。逐技能授權更安全。
它能偵測惡意技能嗎？
它可以標示危險的命令模式，但不能取代完整的安全審查。
它會碰觸哪些檔案？
文件提到 ~/.claude/settings.json，以及 ~/.claude/skills 下的已安裝技能檔案。
新權限何時會套用？
該技能指出，權限更新會在新的工作階段或執行 /clear 後套用。
開發者詳細資訊
作者
guo-yu授權
MIT
Skillstore 修訂
r1
版本提示
作者未聲明版本。
引用
1ffa7643651792ccb4bd3b15d924d2c97edff755
維護新鮮度
2026/7/18
使用量
5 次下載 · 275 次瀏覽
檔案結構