Scan Report
5 /100
scienceclaw-investigate
Run a multi-agent autonomous scientific investigation on any topic
This skill is a documentation-only wrapper that invokes an external scientific research tool. No malicious code, obfuscation, or hidden functionality detected in the SKILL.md.
Safe to install
The skill is safe to use. However, verify the integrity of the external scienceclaw installation at ~/scienceclaw before running, as this skill relies on that external tool.
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Shell | WRITE | WRITE | ✓ Aligned | SKILL.md:26 bash command invocation |
| Filesystem | NONE | NONE | — | No file operations declared or observed |
| Network | NONE | NONE | — | Network behavior delegated to external scienceclaw tool |
| Environment | READ | READ | ✓ Aligned | SKILL.md:6 requires ANTHROPIC_API_KEY |
File Tree
1 files · 4.6 KB · 97 lines Markdown 1f · 97L
└─
SKILL.md
Markdown
Security Positives
✓ Skill is documentation-only with no embedded executable code
✓ Shell execution is explicitly declared and necessary for functionality
✓ No base64, obfuscation, or anti-analysis patterns detected
✓ No credential harvesting beyond required API key usage
✓ No suspicious network patterns or C2 indicators
✓ No file system access beyond standard directory navigation
✓ No supply chain risks since no dependencies in the skill package