K Deep Research

• Analyzed: May 11, 2026, 4:51 AM
• Findings: 1
• Version: 2.0.1
• Hash: abcccba8e28160f88dad16b65c59f1d19c60b71938e97813c4db438690447d23
• Source: palm
• External report: View on VirusTotal

Findings (1)

The OpenClaw AgentSkills skill bundle 'k-deep-research' is classified as benign. Its primary purpose is to guide an AI agent through a systematic deep research methodology. The skill's extensive documentation (SKILL.md and reference files) provides detailed instructions for the agent on sourcing, credibility scoring, pattern recognition, and even adversarial analysis, all aimed at improving research quality. While the skill uses strong prompt-injection phrases (e.g., 'you MUST', 'DO NOT skip') to enforce its methodology, these are designed to ensure thorough research, not to subvert the agent for malicious purposes like data exfiltration or unauthorized actions. The `scripts/validate.sh` file includes a security check for secrets, which is a positive indicator. Network access via `curl` or `wget` is required, but this is plausible and necessary for web-based research. There is no evidence of intentional harmful behavior, credential theft, data exfiltration, persistence, or obfuscation.