registry / @sema-agent/core / 1.307.0

@sema-agent/core@1.307.0

Stateless, task-oriented AI agent core

AI Security Review

scanned 1d ago · by lpm-firewall-ai

No confirmed malicious attack surface. The package is an agent-core library and an explicit `sema-tb` CLI that can invoke configured model providers and agent tools at runtime.

Static reason

High-risk behavior combination matched malicious policy.

Trigger

Explicit import/API use or user invocation of `sema-tb`; no install-time trigger found.

Impact

Powerful runtime capabilities are package-aligned and not silently activated on installation or import.

Mechanism

User-configured AI-agent orchestration with optional shell and web tools.

Rationale

The scanner's reverse-shell and eval findings are false positives from a security-monitor prompt asset and denylist code. Source inspection found no install/import-time execution or concrete malicious chain; runtime capabilities require explicit library/CLI use and are consistent with the package's stated agent-core purpose.

Evidence

package.jsondist/index.jsdist/bin/sema-tb.jsdist/core/auto-mode-prompt-assets.jsdist/core/auto-mode-prompt.jsdist/core/auto-promote.jsdist/tools/web.jsdist/engine/execution-env/node-execution-env.jsdist/core/memory-engine/layout.js

Decision evidence

public snapshot

@sema-agent/core@1.307.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code