@rallycry/conveyor-agent@11.1.3
Conveyor Agent Runner v10 - PTY harness for the task chat (SDK harness for audit/project-chat). Agent-as-User architecture with BaseService patterns. Works locally too.
AI Security Review
scanned 3d ago · by lpm-firewall-aiThe high-risk scanner labels did not establish malicious behavior in the inspected source. The publish hook is build-only, credential handling stays within workspace paths, and the agent settings are created for an invoked PTY session with a plan-mode gate.
Decision evidence
public snapshot- package.json exposes the agent commands and sets prepublishOnly to the build command `tsup`.
- dist/boot-3FWYGHSN.js copies an existing workspace Claude credential file to a pod-local path, then relinks it; boot setup calls this helper.
- dist/chunk-V5WR2F2P.js fetches a missing `.claude/grimoire` submodule only when the workspace declares it and a GitHub token is available.
- dist/chunk-VXGZSCRL.js stores transcripts under the Claude project path and creates hook settings in a temporary `conveyor-pty-` session directory.
- The PTY settings auto-approve tool calls, while plan mode retains its tool gate; the caller enables that gate when permission mode is `plan`.
- The inspected connection subscriptions, query-cache focus/online handlers, and watchdog configuration show ordinary runtime behavior.
Source & flagged code
8 flagged · loading sourcePackage source references child process execution.
dist/chunk-VXGZSCRL.jsView on unpkg · L3798Source spawns a local helper that also contains network and dynamic execution context; review data flow before blocking.
dist/chunk-VXGZSCRL.jsView on unpkg · L415Findings
Affected versions and remediation
This report applies to @rallycry/conveyor-agent@11.1.3.
See all versions of @rallycry/conveyor-agent for other recorded verdicts.
- Review the evidence and your use of @rallycry/conveyor-agent@11.1.3 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.