中文 EN

Scan Report

Scan New Files

⚠ Low Risk — Risk Score 15/100
Last scan:2 days ago Rescan
15 /100
A股财务分析
通过调用Prana平台上的远程agent完成A股上市公司财务状况分析，从6个维度展示，生成交互式HTML报告
A legitimate A-share financial analysis skill with thin, transparent client scripts that properly declare all network calls and environment variable usage, featuring strong user-confirmation guardrails.
Skill NameA股财务分析
Duration31.9s
Enginepi
✓
Safe to install
This skill is safe to use. No malicious patterns detected. The extensive documentation could be simplified but represents best-practice user protection rather than a security risk.

Findings 1 items

Severity Finding Location
Low
Overly verbose documentation with minor clarity issues
SKILL.md is 223 lines with repeated 'Confirmation Required: NO/YES' flags and nearly identical rules restated multiple times. The rules around Step 2 subprocess/curl usage are documented (correctly labeled as NO/SHELL:WRITE), but the sheer length may obscure important distinctions. No security impact, but reduces auditability.
Step 2 硬性要求（无论上下文如何均适用）...
→ Consider condensing repetitive process rules while preserving the critical security guardrails (user confirmation before API key fetch, no key overwriting without permission).
 SKILL.md:1
ResourceDeclaredInferredStatusEvidence
Filesystem NONE NONE — No file system access in any script
Network READ/WRITE READ/WRITE ✓ Aligned POST to claw-uat.ebonex.io declared in SKILL.md and skill.json
Shell WRITE WRITE ✓ Aligned env|grep and curl documented in SKILL.md; scripts do not execute shell
Environment READ READ ✓ Aligned PRANA_SKILL_API_FLAG read declared in skill.json and SKILL.md
Skill Invoke WRITE WRITE ✓ Aligned POST /api/claw/agent-run invokes remote skill
Clipboard NONE NONE — No clipboard access
Browser NONE NONE — No browser access
Database NONE NONE — No database access
7 findings
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/api/v2/api-keys
SKILL.md:15
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/api/claw/agent-run
SKILL.md:17
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/api/claw/agent-result
SKILL.md:19
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/api/claw/skill-purchase-history-url
SKILL.md:21
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/skill-purchase-history-url?pay_token=xxxxxxx
SKILL.md:199
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io
SKILL.md:218
🔗
Medium External URL 外部 URL
https://claw-uat.ebonex.io/
SKILL.md:221

File Tree

4 files · 28.8 KB · 604 lines
Markdown 1f · 223L JavaScript 1f · 163L Python 1f · 163L JSON 1f · 55L
├─ ▾ 📁 scripts
│ ├─ 📜 prana_skill_client.js JavaScript 163L · 5.1 KB
│ └─ 🐍 prana_skill_client.py Python 163L · 5.7 KB
├─ 📋 skill.json JSON 55L · 2.0 KB
└─ 📝 SKILL.md Markdown 223L · 15.9 KB

Security Positives

✓ All network requests go to a single, clearly declared domain: claw-uat.ebonex.io
✓ No base64, eval, or obfuscated code in either client script
✓ API key is read from environment variable only; never hardcoded or logged
✓ Strong user-confirmation gate before fetching API keys (Step 2 requires explicit user consent)
✓ Scripts are thin wrappers (~163 lines each) with straightforward, auditable logic
✓ No access to sensitive paths (~/.ssh, ~/.aws, .env, etc.)
✓ No credential exfiltration or data theft patterns
✓ No remote script execution (curl|bash, wget|sh)
✓ skill.json and SKILL.md declare all network requests and environment variables
✓ No subprocess, os.system, or shell=True usage in client scripts
✓ Both Node.js and Python clients are functionally identical and independently verifiable