gm-plugkit@2.0.2505
Bootstrap and daemon-spawn tool for gm plugkit binary. Downloads the correct platform wasm, verifies SHA256, and launches agentplug-runner (the native wasm host) as the spool watcher daemon.
AI Security Review
scanned 3mo ago · by lpm-firewall-aiThis package establishes a broad, remotely mutable AI-agent instruction and daemon surface when explicitly run. The absence of lifecycle hooks prevents a malicious/install-block verdict under the stated policy, but the capability warrants warning.
Decision evidence
public snapshot- bootstrap.js:803 invokes project wiring on normal CLI boot.
- bootstrap-shared.js:266-289 creates/rewrites CLAUDE.md and AGENTS.md to load .gm instructions.
- bootstrap.js:624-704 installs bundled or remotely discovered skills into ~/.agents/skills and ~/.claude/skills.
- bootstrap.js:610-621 fetches arbitrary skill names/content from AnEntrypoint/gm main branch.
- bootstrap.js:931-961 downloads and installs a native agentplug-runner, then bootstrap.js:968-1007 detaches it as a daemon.
- SKILL.md directs agents to follow a live instruction stream and forbids normal completion.
- package.json has no preinstall, install, or postinstall hook.
- The behavior is reached through the explicit gm-plugkit CLI, not package import.
- Downloaded runner and wasm paths have SHA-256 checks, though trust remains remote-controlled.
Source & flagged code
3 flagged · loading sourcePackage source references child process execution.
bootstrap-shared.jsView on unpkg · L5Source downloads or fetches remote code and executes it.
bootstrap.jsView on unpkg · L7Findings
Affected versions and remediation
This report applies to gm-plugkit@2.0.2505.
See all versions of gm-plugkit for other recorded verdicts.
- Review the evidence and your use of gm-plugkit@2.0.2505 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.