Skill AGI
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill is coherent and benign, but it intentionally creates persistent learning notes that can influence future sessions.
This skill has no code or credentials and is designed to keep a small local list of lessons. Install it if you want that persistent memory, but periodically inspect or delete entries, avoid secrets, and approve any promotion into broader agent memory files.
Findings (1)
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 carry forward a mistaken lesson or project detail and apply it later without the full original context.
The skill intentionally stores learnings and reuses or promotes them as future agent context, so inaccurate or sensitive entries could influence later sessions.
Before starting a complex task, read `learnings.md`... Promote to workspace `MEMORY.md`, `TOOLS.md`, or `AGENTS.md`
Review learnings.md periodically, avoid recording secrets or unverified instructions, and only promote entries to durable memory/config files when you trust them.