Back to skill
Skillv1.0.12

ClawScan security

Gws Keep · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

SuspiciousMar 31, 2026, 6:34 PM
Verdict
suspicious
Confidence
medium
Model
gpt-5-mini
Summary
The skill's declared purpose (manage Google Keep via the gws CLI) is plausible, but the runtime instructions reference an external shared SKILL.md and a generate step that can create auth/config outside the skill without having declared those config paths or credentials — this mismatch is concerning.
Guidance
Before installing or enabling this skill: (1) verify the 'gws' CLI you will use is from a trusted source and inspect its behavior; (2) locate and manually inspect the referenced ../gws-shared/SKILL.md (or equivalent shared config) to see what auth or secrets it contains — do not let the agent auto-create it without review; (3) avoid running 'gws generate-skills' unless you trust the gws binary and understand what files it will write and what credentials it will store; (4) prefer testing in an isolated environment or container first; (5) if you cannot find the shared config or a trustworthy source for gws, treat this skill as potentially risky because it implicitly requires credential/config access that isn't declared.

Review Dimensions

Purpose & Capability
noteRequiring the 'gws' CLI matches the stated purpose (Google Keep management). However the SKILL.md depends on a separate ../gws-shared/SKILL.md for auth and global flags, which is not declared in the skill's requirements (no config paths or env vars). That external dependency is not explained in the registry metadata.
Instruction Scope
concernThe SKILL.md explicitly tells the agent to read ../gws-shared/SKILL.md for auth and security rules, and if missing to run 'gws generate-skills' to create it. That directs the agent to read a file outside the skill bundle and to create/configure files on disk via the gws binary — actions outside the skill's declared scope and not enumerated in requires.configPaths.
Install Mechanism
okThis is an instruction-only skill with no install spec, so nothing is written by the skill itself. The model expects a preinstalled 'gws' binary; lack of an install step lowers installer risk but places trust on that external binary.
Credentials
concernNo environment variables or credential fields are declared, yet the skill points to a shared SKILL.md that presumably contains auth/global flags. That implies access to credentials or config files that are not declared here, which is disproportionate to the metadata shown.
Persistence & Privilege
okThe skill does not request always:true and has no install actions. It can be invoked autonomously (default), which is normal; there is no evidence it modifies other skills or system-wide agent settings.