registry / nanoclaw-agenttrace / 0.2.0

nanoclaw-agenttrace@0.2.0

Live agent activity traces for NanoClaw — tools, tasks, reasoning summaries, and silence keepalives

AI Security Review

scanned 3h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. No unconsented install-time attack surface is established. An explicit CLI install modifies a NanoClaw fork to add agent-activity telemetry and a first-party Claude skill; emitted activity is delivered through the fork's existing channel adapters.

Static reason

One or more suspicious static signals were detected.

Trigger

User runs `nanoclaw-agenttrace install`, `upgrade`, or `sync-skill` in a NanoClaw project.

Impact

The installed NanoClaw project can expose summarized reasoning, partial output, and optionally redacted tool data to its configured activity channel.

Mechanism

Explicit project patching and agent activity telemetry injection.

Rationale

This is an explicit-user-command, first-party NanoClaw extension installer rather than concrete malware. Warn because it performs broad project and agent-surface mutation and can publish agent telemetry, including more sensitive content in opt-in `trace_full` mode.

Evidence

package.jsondist/cli/bin.jsdist/cli/install.jsdist/cli/patch.jsskills/add-agenttrace/resources/agenttrace-boot.tsskills/add-agenttrace/resources/agenttrace-delivery.tsskills/add-agenttrace/resources/runner/observe.tsskills/add-agenttrace/resources/runner/writer.tsskills/add-agenttrace/resources/runner/sanitize.tssrc/agenttrace-*.tssrc/index.ts

nanoclaw-agenttrace@0.2.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code