Roundtable

Multi-agent debate council — spawns 3 specialized sub-agents in parallel (Scholar, Engineer, Muse) for Round 1, then optional Round 2 cross-examination to ch...

This skill appears to do what it says and does not request credentials or install code. Before installing or running it, consider: 1) Logging: the setup defaults include session logging to memory/roundtable/. If you or your users may submit sensitive data, choose 'No logging' during setup or ensure the workspace memory storage meets your privacy requirements. 2) Config file: the wizard writes config.json into the skill directory after confirmation — review its contents before enabling features like round2 or max_budget. 3) Web/tooling: the Scholar role mentions web results; if the host agent has browsing or external connectors, those tools and their credentials are outside this skill — ensure those integrations are controlled. 4) Cost and data exposure: multi-agent runs multiply model calls; avoid sending secrets in queries. If these items are acceptable, the skill is coherent and can be used; if you need stronger guarantees, disable logging and review the saved config/logs after setup.

Type: OpenClaw Skill Name: roundtable Version: 0.4.1 The OpenClaw 'roundtable' skill bundle is classified as benign due to its explicit and robust security measures against common attack vectors. Both `SKILL.md` and `README.md` contain mandatory instructions for the AI agent to treat user queries and web search results as 'UNTRUSTED INPUT,' explicitly forbidding execution of embedded instructions, role changes, or unauthorized file/network access. Furthermore, the skill enforces a fixed logging path (`memory/roundtable/`) to prevent path traversal attacks and explicitly instructs the agent to 'Never include secrets/API keys' in logs. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or obfuscation. The skill actively mitigates vulnerabilities rather than introducing them.