Skill Shell
ReviewAudited by ClawScan on May 10, 2026.
Overview
This is mostly a cautious, instruction-only skill reviewer, but it also tells the agent to copy ideas from untrusted skills into persistent workflow and memory files without clearly requiring user approval.
This skill appears safe for read-only review of other skills, but do not let it automatically update persistent workflow or memory files. If you use its “absorb ideas” path, require explicit approval and inspect the exact changes before saving them.
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.
A malicious or poorly designed external skill could have its instructions summarized into persistent local guidance, affecting future sessions even if the original skill was not installed.
These are persistent workflow or memory files that can influence future agent behavior. The instruction is purpose-related, but it does not clearly require user approval, source labeling, sanitization, or review before importing ideas from an external skill.
Instead, absorb the useful parts into the local workflow: - behavioral patterns -> `SOUL.md` - workflow rules -> `AGENTS.md` - tool gotchas -> `TOOLS.md` - session-specific lessons -> `memory/YYYY-MM-DD.md`
Use this skill only for read-only evaluation unless the user explicitly approves any persistent edits. Review diffs before writing to SOUL.md, AGENTS.md, TOOLS.md, or memory files, and store neutral summaries rather than copying instructions wholesale.