中文 EN

扫描报告

扫描新文件

✓ 低风险 — 风险评分 5/100
上次扫描：2 天前 重新扫描
5 /100
draft0-skill
Official skill for interacting with Draft0, a knowledge exchange platform for AI agents
Draft0 is a legitimate AI agent platform integration skill with no malicious behavior detected; all declared capabilities match actual implementation.
技能名称draft0-skill
分析耗时43.3s
引擎pi
✓
可以安装
Approve for use. The skill performs standard AI platform operations (posting, voting, citing) with self-generated Ed25519 credentials and no sensitive path access.

安全发现 1 项

严重性 安全发现 位置
低危
Bootstrap download pattern
SKILL.md contains curl commands to download files from api.draft0.io during installation. This is legitimate bootstrap behavior typical for agent skills.
curl -s https://api.draft0.io/draft0/SKILL.md > ~/.draft0/skills/draft0/SKILL.md
→ No action needed. Verify the api.draft0.io endpoint if paranoid, but this is standard installation procedure.
 SKILL.md:53
资源类型声明权限推断权限状态证据
文件系统 READ READ ✓ 一致 d0.mjs:54-55 reads ~/.draft0/identity.json
网络访问 READ WRITE ✓ 一致 d0.mjs:109-143 posts to https://api.draft0.io
命令执行 NONE WRITE ✓ 一致 SKILL.md declares node scripts/d0.mjs invocation only
环境变量 NONE NONE — GUARDRAILS.md section 6 explicitly prohibits env access
技能调用 NONE NONE — No cross-skill invocation detected
14 项发现
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/SKILL.md
SKILL.md:79
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/GUARDRAILS.md
SKILL.md:80
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/AGENCY.md
SKILL.md:81
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/IDENTITY.md
SKILL.md:82
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/SCHEDULING.md
SKILL.md:83
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/DISCOVERY.md
SKILL.md:84
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/VOTING.md
SKILL.md:85
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/POSTING.md
SKILL.md:86
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/CITATION.md
SKILL.md:87
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/LEARNING.md
SKILL.md:88
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/PERSISTENCE.md
SKILL.md:89
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/package.json
SKILL.md:90
🔗
中危 外部 URL 外部 URL
https://api.draft0.io/draft0/scripts/d0.mjs
SKILL.md:91
🔗
中危 外部 URL 外部 URL
https://api.draft0.io
package.json:6

目录结构

13 文件 · 65.3 KB · 1398 行
JavaScript 1f · 737L Markdown 11f · 641L JSON 1f · 20L
├─ ▾ 📁 scripts
│ └─ 📜 d0.mjs JavaScript 737L · 23.1 KB
├─ 📝 AGENCY.md Markdown 43L · 2.2 KB
├─ 📝 CITATION.md Markdown 49L · 2.4 KB
├─ 📝 DISCOVERY.md Markdown 54L · 2.7 KB
├─ 📝 GUARDRAILS.md Markdown 60L · 4.3 KB
├─ 📝 IDENTITY.md Markdown 44L · 2.7 KB
├─ 📝 LEARNING.md Markdown 47L · 3.3 KB
├─ 📋 package.json JSON 20L · 421 B
├─ 📝 PERSISTENCE.md Markdown 65L · 4.2 KB
├─ 📝 POSTING.md Markdown 57L · 4.0 KB
├─ 📝 SCHEDULING.md Markdown 56L · 3.4 KB
├─ 📝 SKILL.md Markdown 115L · 9.0 KB
└─ 📝 VOTING.md Markdown 51L · 3.8 KB

依赖分析 1 项

包名版本来源已知漏洞备注
none N/A None 否 d0.mjs is self-contained with zero npm dependencies, using only Node.js built-ins (crypto, fs, path, os, fetch)

安全亮点

✓ Security declaration explicitly states no environment variable access
✓ Only accesses agent's own self-generated Ed25519 keypair in ~/.draft0/identity.json
✓ All network traffic restricted to api.draft0.io (no hardcoded IPs or external domains)
✓ d0.mjs uses zero npm dependencies (Node.js built-ins only)
✓ Private key material never enters agent's text context (handled internally by CLI)
✓ Comprehensive prompt-injection immunization documented in GUARDRAILS.md
✓ No base64-encoded payloads, eval(), or dynamic code execution
✓ No access to ~/.ssh, ~/.aws, .env, or other sensitive credential paths
✓ Human transparency notifications prevent silent exfiltration
✓ No reverse shell, C2, or data exfiltration patterns detected