AI大佬思想蒸馏框架
AdvisoryAudited by VirusTotal on Apr 21, 2026.
Overview
Type: OpenClaw Skill Name: thought-leader-forum Version: 3.0.0 The 'thought-leader-forum' skill bundle is a specialized simulation tool designed to model expert opinions from the AI, hospitality, and venture capital sectors. The bundle consists of structured markdown files defining various personas (e.g., Elon Musk, Sam Altman) and background guides for a specific corporate AI strategy. Analysis of SKILL.md and the associated shell script (run-discussion.sh) reveals no malicious intent, data exfiltration, or unauthorized execution; the instructions are strictly focused on generating labeled AI simulations and maintaining legal compliance through mandatory disclaimers.
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.
A user could misunderstand or share simulated statements as if they were real quotes if the required labels or disclaimers are removed.
The skill intentionally generates simulated viewpoints of named real people. The artifacts clearly disclose this and require labels, which mitigates the trust risk, but users still need to avoid treating outputs as authentic statements.
所有输出必须标注【观点模拟：XXX人物名】...不声称代表任何真实人物的立场、观点或言论。
Keep the 【观点模拟】 labels in all outputs, do not present generated content as authorized views, and verify important business or legal claims independently.
If web/search tools are enabled, the agent may use external sources to shape the discussion output, which can introduce inaccurate or unreviewed context.
This instructs the assistant to supplement missing background by searching if tools are available. That is aligned with the public-research persona purpose, but it is still external tool use the user should notice.
有没有具体的背景资料？（没有的话我会根据议题自行搜索补充）
Ask the assistant to cite sources or avoid web search when you want the discussion limited to provided materials.
Users have less external provenance information for deciding whether to trust the skill’s author and content.
The registry metadata does not provide a verifiable source or homepage. The reviewed artifacts do not show risky dependencies or remote installers, so this is only a provenance note.
Source: unknown; Homepage: none
Review the included markdown and script contents before use, especially if using the skill for business decision support.