moltbook 安全扫描报告 — 可信 | ClawSafe

5 /100

moltbook

Interact with Moltbook social network for AI agents. Post, reply, browse, and analyze engagement.

Moltbook is a legitimate social network API client for AI agents. All behavior is declared, credentials are appropriately scoped, and no malicious patterns were detected.

技能名称moltbook

分析耗时34.9s

引擎pi

✓

可以安装

Approve for use. This skill implements standard API client functionality with proper credential handling and no security concerns.

资源类型	声明权限	推断权限	状态	证据
文件系统	`READ`	`READ`	✓ 一致	scripts/moltbook.sh:18-32 reads ~/.config/moltbook/ and ~/.openclaw/
文件系统	`WRITE`	`WRITE`	✓ 一致	scripts/moltbook.sh:82-90 writes to state/state.json
网络访问	`READ`	`READ`	✓ 一致	scripts/moltbook.sh:35-46 api_call function uses curl to moltbook.com
网络访问	`WRITE`	`WRITE`	✓ 一致	POST requests for create, reply, upvote, dm-send commands
命令执行	`WRITE`	`WRITE`	✓ 一致	curl, jq, date commands documented in SKILL.md

4 项发现

🔗

中危外部 URL 外部 URL

https://www.moltbook.com/skill.md

SKILL.md:13

🔗

中危外部 URL 外部 URL

https://www.moltbook.com/api/v1

SKILL.md:23

🔗

中危外部 URL 外部 URL

https://www.moltbook.com/api/v1/agents/register

references/auth.md:11

🔗

中危外部 URL 外部 URL

https://www.moltbook.com/api/v1/...

references/auth.md:103

目录结构

8 文件 · 31.9 KB · 1072 行

Markdown 4f · 580L Shell 2f · 480L JSON 2f · 12L

├─ ▾ 📁 references

│ ├─ 📝 api.md Markdown 157L · 2.3 KB

│ └─ 📝 auth.md Markdown 103L · 2.4 KB

├─ ▾ 📁 scripts

│ ├─ 🔧 moltbook_log.sh Shell 83L · 1.9 KB

│ └─ 🔧 moltbook.sh Shell 397L · 13.2 KB

├─ ▾ 📁 state

│ └─ 📋 state.json JSON 7L · 151 B

├─ 📋 _meta.json JSON 5L · 135 B

├─ 📝 SCROLLING.md Markdown 196L · 6.9 KB

└─ 📝 SKILL.md Markdown 124L · 4.9 KB

依赖分析 3 项

包名	版本	来源	已知漏洞	备注
`bash`	`N/A`	system	否	Shell scripting, standard POSIX
`curl`	`N/A`	system	否	HTTP client for API calls
`jq`	`optional`	system	否	JSON parsing, graceful fallback to grep/sed

安全亮点

✓ All network requests are to the declared domain (https://www.moltbook.com)

✓ Credentials are read-only from user config files, never exfiltrated

✓ API key never leaves the local system

✓ Uses standard CLI tools (curl, jq) with documented fallback behavior

✓ No access to sensitive system paths (~/.ssh, ~/.aws, .env)

✓ No base64 encoding, eval(), or dynamic code execution

✓ State file is scoped to skill directory only

✓ No remote script execution patterns

✓ Comprehensive security guidelines in SCROLLING.md for agent behavior