registry / inferock-bench / 0.1.4

inferock-bench@0.1.4

Local LLM billing proxy that turns OpenAI and Anthropic calls into per-call cost and failure receipts.

AI Security Review

scanned 3h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. An explicit agent-mode test can install and run a third-party coding-agent binary. No install-time execution or foreign AI-agent configuration mutation is present.

Static reason

No blocking static signals were detected.; previous stored version diff introduced dangerous source

Trigger

User runs `inferock-bench test --generator agent` and confirms the agent-install consent.

Impact

The downloaded agent receives edit/read/bash capability inside the selected workspace; this is a guarded extension lifecycle risk.

Mechanism

Consent-gated download, unpack, configuration, and execution of a local OpenCode agent.

Rationale

This is not concrete malicious behavior, but it provisions and launches a capable third-party coding agent after explicit consent. Policy classifies this first-party, guarded agent-extension setup as warn rather than block.

Evidence

package.jsondist/cli.jsdist/agent-mode/provisioner.jsdist/agent-mode/opencode-adapter.jsdist/coverage-suite/agent-runner.jsdist/telemetry.jsopencode.jsoninstall-manifest.json

Network endpoints4

registry.npmjs.org/opencode-ai/-/opencode-ai-1.17.13.tgzregistry.npmjs.org/opencode-linux-x64/-/opencode-linux-x64-1.17.13.tgz

inferock-bench@0.1.4

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code