Game Theory
PassAudited by VirusTotal on May 12, 2026.
Overview
Type: OpenClaw Skill Name: game-theory Version: 1.0.0 The skill bundle consists entirely of markdown documentation and analytical frameworks related to game theory in crypto. While it extensively details various attack vectors and adversarial behaviors (e.g., flash loan attacks, sandwich attacks, governance capture) in files like `references/governance-attacks.md` and `references/mev-strategies.md`, this content is presented purely for analytical understanding and defense, aligning with the skill's stated purpose of 'predicting adversarial behavior' and 'identifying attack vectors.' The `SKILL.md` file contains no prompt injection attempts instructing the agent to perform unauthorized actions, exfiltrate data, or execute malicious code. No executable code or high-risk behaviors were found.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The skill may help analyze how governance attacks or social manipulation work, which is useful defensively but could be misused if prompted for real-world attack planning.
The reference material includes explicit attacker-oriented governance manipulation tactics. It is framed as being for defensive understanding and fits the skill’s stated purpose, but it is dual-use content that users should keep bounded.
### For Attackers (to understand defense) 1. **Accumulate quietly:** Avoid detection ... 4. **Use social engineering:** Control narrative 5. **Extract incrementally:** Many small raids vs. one big one
Use the skill for defensive analysis, mechanism design, and risk review. Avoid using it to plan deception, governance manipulation, or exploit execution against real communities or protocols.