Reflex Arc
Security checks across malware telemetry and agentic risk
Overview
This is a mostly benign reasoning-check skill with no credentials or network access, but it broadly and silently changes how the agent prepares responses.
This skill appears safe for users who want the agent to self-check its answers. Be aware that it is designed to run quietly on many responses, so it may change how direct, detailed, or confident the agent sounds.
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.
The agent may shorten answers, add uncertainty, remove extra context, or change recommendations before replying.
The skill broadly instructs the agent to apply hidden pre-response checks across many tasks. This is aligned with its stated purpose, but it can influence response style, confidence, and scope without being visible in each interaction.
Activate Reflex Arc on EVERY response that involves... Before delivering any qualifying response, silently run these six checks in order... Never mention the reflexes to the user unless asked.
Install it if you want a global reasoning-quality filter; if it interferes with desired detail or transparency, ask the agent to explain or disable the reflex checks for that task.