agentix-cli@0.145.0
The self-hosted agent harness for small teams. Runs a team of AI agents across Telegram, WhatsApp, GitLab, GitHub, crons, webhooks and mesh tasks, on Claude Code, Codex, OpenCode or any LLM API. Web wizard for operators, CLI for engineers.
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe source shows automatic first-party skill setup during daemon startup in registered workspaces, so residual agent extension lifecycle risk remains. The inspected credential request goes to Anthropic, and the install hook performs a conditional SQLite rebuild; neither establishes a separate attack.
Decision evidence
public snapshot- Daemon startup calls a routine that installs or updates the package’s memory skill in each registered agent workspace.
- The routine creates the workspace’s `.claude/skills` directory and writes `remember.md`; it also syncs agent memory into the workspace.
- Agent setup writes a workspace `CLAUDE.md` from configured personality data when that file is absent. This is first-party agent configuration setup with no consent check shown in the inspected path.
- The postinstall hook is declared in the manifest and its script checks the local SQLite binding, rebuilding the declared dependency only after an ABI mismatch.
- Discovered Anthropic credentials are sent to the Anthropic messages endpoint; the inspected request does not show a different recipient.
- The inspected configuration writes target registered agent workspaces; no global agent settings mutation or unrelated credential recipient was established.
- The postinstall script’s inspected behavior is a conditional native dependency rebuild, not a consumer approval-policy change.
Source & flagged code
20 flagged · loading sourcePackage defines install-time lifecycle scripts.
package.jsonView on unpkgInstall-time lifecycle script is not statically allowlisted and needs review.
package.jsonView on unpkgFindings
Affected versions and remediation
This report applies to agentix-cli@0.145.0.
See all versions of agentix-cli for other recorded verdicts.
- Review the evidence and your use of agentix-cli@0.145.0 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.