Back to skill
Skillv1.0.0
VirusTotal security
Claude Api Anthropic · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
ReviewApr 30, 2026, 6:18 AM
- Hash
- d160f4789fe42d6522f179b470e2ea91791040a27f20bb8825f1ee85de8ff4d7
- Source
- palm
- Verdict
- suspicious
- Code Insight
- Type: OpenClaw Skill Name: claude-api-anthropic Version: 1.0.0 The bundle provides documentation for non-existent Anthropic models (e.g., 'claude-opus-4-6') and future-dated API features (dated 2026). It contains forceful instructions in SKILL.md designed to override user preferences regarding model selection and cost, which constitutes a prompt injection risk. While it describes high-risk capabilities like shell execution ('Bash' tool) and permission bypassing ('bypassPermissions') in the agent-sdk files, there is no clear evidence of an intentional payload for data exfiltration or unauthorized access. The primary risk is the agent being misled into using invalid API parameters and potentially dangerous execution modes based on deceptive documentation.
- External report
- View on VirusTotal