migrate-skill-agent

SUSPICIOUS: the stated purpose matches the behavior, but the skill updates an unspecified `skills-store` repo and installs its contents into active agent skill directories, creating a medium supply-chain and high transitive-trust risk. No direct credential harvesting or overt malware behavior is shown in the provided text.

This module is primarily a high-impact installer/synchronizer for repository assets. It does not show direct malware primitives in the fragment, but it carries substantial supply-chain and operational risk: it performs an unpinned/unverified `git pull` and then copies/syncs or symlinks repository content into $HOME/$PWD using potentially destructive operations (`rm -rf`, `rsync --delete`, and symlink replacement). Verification is existence-only, so malicious/tampered content that still matches expected paths would not be detected here. Exact exploitability depends on unseen input validation and how $STORE/find_store and upstream argument parsing constrain $NAME/$TYPE/$FLAVOR/$LINK.