power-engineer

This fragment describes a legitimate project configuration utility, not an evident malware payload. However, it introduces persistent shell hooks and aggressive automatic memory collection. The generated rules may store sensitive project facts such as credential locations, endpoints, contacts, and authentication details without explicit user confirmation. Review and constrain the memory categories, inspect the omitted hook scripts before installation, ensure generated state is gitignored, and verify that hook commands cannot be replaced or traversed through untrusted paths. The supplied fragment contains no demonstrated exfiltration or destructive behavior.

The fragment is an installation-plan document, not directly malicious code. It presents a meaningful supply-chain risk because it instructs users to run many unpinned npx commands from third-party sources using @latest and automatic confirmation. Any installed package can execute installation or runtime code, so provenance review, version pinning, lockfiles, isolated execution, and maintainer verification are recommended. No direct credential theft, exfiltration, persistence, or destructive behavior is visible in this fragment.

SUSPICIOUS. The visible skill is coherent with a project-setup purpose, but its trust-critical behavior is deferred to hidden flow files and it is explicitly designed to install additional skills, creating a transitive trust chain that cannot be reviewed here. No direct malware or credential exfiltration is visible in the excerpt, but the missing installer details make the overall risk medium-high.

This is an installer specification rather than an observed malware payload. Its intended behavior is to install remote skills and plugins, but it presents a significant supply-chain security risk because it executes npx commands that fetch unverified code and inconsistently permits the mutable @latest version. The version-pin contradiction should be corrected, remote sources should be allowlisted and integrity-verified, command arguments safely validated, and permission-rule cleanup enforced before use.

The supplied content is an explicit permission-automation specification, not direct malware. It includes consent and cleanup controls, but it creates an auto-approval hook and a wildcard permission for `npx skills@latest` commands, with a fragile shell-based parser and optional persistence. This presents a moderate security risk because a compromised or unexpected command could receive approval, although there is no evidence of data theft, system damage, or other malicious payload in the fragment.