registry / @sema-agent/core / 1.328.0

@sema-agent/core@1.328.0

Stateless, task-oriented AI agent core

AI Security Review

scanned 3h ago · by lpm-firewall-ai

No confirmed malicious attack surface was established. The package exposes AI-agent and workflow utilities, including explicitly supplied workflow validation, but inspected code does not show covert execution or exfiltration.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Consumer explicitly invokes agent/workflow functionality or the packaged build lifecycle runs.

Impact

No concrete unconsented payload, persistence, credential theft, or network exfiltration path found.

Mechanism

Agent-core utilities with guarded workflow validation and host-provided execution abstractions.

Rationale

The scanner’s critical reverse-shell signal is attributable to embedded security-monitor prompt text, while the inspected lifecycle and implementation files show no concrete malicious chain. Dynamic workflow handling is package-aligned and guarded rather than covert payload execution.

Evidence

package.jsondist/core/auto-mode-prompt-assets.jsdist/core/auto-promote.jsdist/core/session-store.jsdist/orchestration/run-workflow-tool.jsdist/orchestration/dev-vm-script-runner.jsdist/engine/execution-env/node-execution-env.jsdist/bin/sema-tb.js

Decision evidence

public snapshot

@sema-agent/core@1.328.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code