@passioncode-ai/fabric-agent-adapter@0.56.0
Create Fabric-compatible agents, run them as always-alive local services with dashboards (fabric-service/0.1), and adapt existing projects to the Fabric Agent Contract. This package is the installer CLI.
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe package has no npm install hook, and its launcher writes skills only when invoked. However, the Claude plugin automatically reports Git activity facts during sessions, with no general opt-out in the package; this is disclosed runtime telemetry that warrants a warning.
Decision evidence
public snapshot- Claude plugin hooks run automatically on session events, including session start.
- The hook records Git facts for each event, including a commit ID and change counts.
- When the Fabric observation tool is available, the worker sends those facts to it.
- The provided opt-out disables transcript-derived repository discovery; it does not disable the ordinary Git facts.
- The npm manifest has no install lifecycle script.
- The observation data is limited to session facts; the code does not send file contents or local paths in the facts payload.
- The hook description discloses the session facts and repository metadata sent to Fabric.
Source & flagged code
6 flaggedManifest-trigger-reachable source writes behavior-bearing configuration into a user or project AI-agent control surface.
bin/fabric-agent-adapter.jsView on unpkgPackage source references dynamic require/import behavior.
bin/fabric-agent-adapter.jsView on unpkg · L14Findings
Affected versions and remediation
This report applies to @passioncode-ai/fabric-agent-adapter@0.56.0.
See all versions of @passioncode-ai/fabric-agent-adapter for other recorded verdicts.
- Review the evidence and your use of @passioncode-ai/fabric-agent-adapter@0.56.0 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.