Prompt Safe

• Analyzed: May 14, 2026, 3:24 AM
• Findings: 1
• Version: 1.0.4
• Hash: b73f8d08755113c4edea3c4bd53be09b5b5eebc482c80b826c2b27619a58eab6
• Source: palm
• External report: View on VirusTotal

Findings (1)

The OpenClaw AgentSkills bundle 'prompt-assemble' is a utility designed for token-safe prompt assembly and memory orchestration for LLM agents. The `SKILL.md` instructions clearly define the skill's purpose and workflow, without any evidence of prompt injection attempts or malicious instructions for the agent. The core implementation in `scripts/prompt_assemble.py` is well-structured, uses local token estimation heuristics, and relies on functions passed as arguments for memory and dialog retrieval, without making any suspicious external network calls, accessing sensitive files, or executing arbitrary code. All components align with the stated goal of preventing token overflow and ensuring API stability, indicating a benign and well-engineered utility.