Trusted — Risk Score 5/100
Last scan:17 hr ago Rescan
5 /100
cnpc
Information assistant for 中国石油 CNPC/PetroChina. Search products, news, financials, and official resources.
This is a simple information assistant skill with only documentation (SKILL.md) - no executable code, scripts, or dependencies. The described behavior is benign: web-based information retrieval about CNPC/PetroChina from official sources.
Skill Namecnpc
Duration33.0s
Enginepi
Safe to install
No action required. This skill contains only documentation with no executable code or suspicious behavior.

Findings 1 items

Severity Finding Location
Low
Web search capability not formally declared Doc Mismatch
SKILL.md states the skill 'uses web search' but does not formally declare network/browser tool permissions in the resource model
This skill uses web search to fetch real-time information about 中国石油 CNPC/PetroChina
→ Consider adding explicit allowedTools declaration if network access is required
SKILL.md:28
ResourceDeclaredInferredStatusEvidence
Filesystem NONE NONE No file access operations in documentation
Network READ READ ✓ Aligned SKILL.md line 28-31: describes web search to fetch info from cnpc.com.cn
Shell NONE NONE No shell command references in documentation
Environment NONE NONE No environment access mentioned
Skill Invoke NONE NONE No inter-skill invocation described
Clipboard NONE NONE No clipboard access mentioned
Browser NONE NONE SKILL.md mentions 'web search' but no explicit browser tool declared
Database NONE NONE No database access mentioned
1 findings
🔗
Medium External URL 外部 URL
https://www.cnpc.com.cn
SKILL.md:16

File Tree

1 files · 2.0 KB · 56 lines
Markdown 1f · 56L
└─ 📝 SKILL.md Markdown 56L · 2.0 KB

Security Positives

✓ No executable code or scripts present - purely documentation
✓ No credential harvesting or environment variable access
✓ No obfuscated code, base64, or suspicious patterns
✓ References only official legitimate sources (cnpc.com.cn)
✓ No supply chain risks (no dependencies)
✓ No persistence mechanisms or backdoors
✓ No prompt injection or jailbreak attempts
✓ No sensitive file path access (~/.ssh, ~/.aws, .env)
✓ Behavior is fully documented in SKILL.md