AGI / Artificial General Intelligence
PassAudited by VirusTotal on May 12, 2026.
Overview
Type: OpenClaw Skill Name: agi Version: 1.0.0 The OpenClaw AGI skill bundle is benign. Its primary purpose is to enhance the AI agent's internal reasoning, planning, and self-awareness, as detailed across `SKILL.md`, `reasoning.md`, and `blindspots.md`. The `SKILL.md` explicitly states strict boundaries: the skill operates within its `~/agi/` directory, never makes network requests, and never accesses files outside its scope, with one exception. The `setup.md` file clarifies this exception: with explicit user consent, it adds a single line to the user's `~/MEMORY.md` (or `~/clawd/MEMORY.md`) for activation. There is no evidence of data exfiltration, malicious execution, persistence mechanisms, or prompt injection designed to subvert the agent for harmful purposes. All instructions are focused on improving the agent's cognitive functions.
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 assistant may remember style or reasoning preferences locally and reuse them later.
The skill stores ongoing observations about the user and interaction patterns, which is purpose-aligned but creates persistent context that can influence future responses.
Over time, you might learn: - User's preferred communication style (verbose vs concise) - Their domain expertise level ... Learn this naturally from interaction. Don't ask.
Review or delete files under ~/agi/ if you do not want this persistent memory, and avoid letting sensitive personal details be saved there.
If approved, this skill may become active automatically in later sessions.
The skill can persist its activation across future sessions by modifying a memory file, but the artifact explicitly says to ask once and require user consent.
"Want me to add AGI to your MEMORY.md so it activates automatically in future sessions?" ... If yes → add ONE line to user's `~/MEMORY.md`
Only approve the MEMORY.md change if you want this behavior by default; remove the added line later if you change your mind.
A user could over-trust the assistant's answers because of the AGI framing.
The branding and capability language may overstate what an instruction-only skill can provide, even though the skill also includes humility and uncertainty guidance.
Activate alongside any other skill to add human-level reasoning, planning, and self-awareness.
Treat this as a reasoning prompt, not actual artificial general intelligence; verify important or high-stakes answers independently.