Multi-Perspective Forum
Security checks across malware telemetry and agentic risk
Overview
The skill is mostly a prompt-only debate tool, but it proposes recurring LLM-powered audits that may collect and send business metrics and logs without clear scoping or approval controls.
Use this skill manually for strategic debates, or only enable its cron workflow if you are comfortable with recurring LLM-powered audits. Before running it, decide exactly which data sources are allowed, redact sensitive logs or customer data, and make sure you know how to disable any scheduled jobs.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
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.
If configured, the skill could continue running audits on a schedule and consuming LLM access after the initial setup.
This instructs use of persistent scheduled LLM-enabled runs. The schedule is disclosed, but the artifact does not define user confirmation, expiry, cleanup, or containment for recurring autonomous activity.
## Cron - Monday 9 AM ET: `0 13 * * 1` (leverage audit) - Friday 9 AM ET: `0 13 * * 5` (retro) Use OpenClaw cron (needs LLM access), not crontab.
Only create the cron jobs intentionally, require explicit confirmation for setup, document how to disable them, and consider adding an expiry or review period.
Sensitive business metrics, lead information, revenue events, or logs could be included in an LLM prompt more broadly than the user intended.
The skill tells the agent to place broad business and operational data into a prompt sent through a spawned model/session, but does not specify redaction, data minimization, allowed sources, retention, or provider data boundaries.
Inject `{DATA}` and `{QUESTION}` into this template ... DATA PAYLOAD: {DATA} ... sessions_spawn: ... model: "github-copilot/claude-opus-4.6" ... Resend analytics, lead pipeline counts, revenue/signups, error logs, outreach totals, cron health, cost burn.Limit inputs to the minimum necessary, redact customer and secret-bearing log data, and require user review before sending audit payloads to the model.