@wayai/cli@0.3.172
WayAI CLI — sync hub configuration between local files and the platform
AI Security Review
scanned 1mo ago · by lpm-firewall-aiThe package does not mutate agent config during npm install, but running the CLI can link or install first-party skills into Claude, OpenCode, and .agents directories. That guarded agent-extension behavior is a warning, not a publish block, because it is user-invoked, same-vendor, and not a hidden lifecycle payload.
Decision evidence
public snapshot- Any non-help wayai command can create .claude/skills and symlink wayai to the existing .agents/skills/wayai skill.
- The authenticated admin skill-install command writes the API file map into .claude, .opencode, and .agents skill directories.
- package.json has no preinstall, install, or postinstall hook; the bin runs only when the user invokes wayai.
- Skill writes are behind an explicit admin command and authenticated API call, and unsafe relative paths are rejected.
- The Claude symlink is skipped for help and version, does nothing if the skill file is already present, and prints what it linked.
- Version refresh fetches skill text only to cache a version string and does not write agent files.
Source & flagged code
7 flaggedManifest-reachable source sends a prompted API credential to a fixed unofficial gateway and persists the redirection.
dist/index.jsView on unpkgFindings
Affected versions and remediation
This report applies to @wayai/cli@0.3.172.
See all versions of @wayai/cli for other recorded verdicts.
- Review the evidence and your use of @wayai/cli@0.3.172 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.