registry / kenship / 0.1.0

kenship@0.1.0

Messaging between people, brokered by their assistants — inside Claude Code.

AI Security Review

scanned 15h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. An explicit `kenship install` configures a user-scope Claude Code extension with remote MCP, hooks, and a statusline. The hooks use the Kenship bearer token and one reads the Claude-provided assistant transcript path to mark delivered messages.

Static reason

No blocking static signals were detected.

Trigger

User runs `kenship install`, then starts or uses Claude Code.

Impact

A third-party service gains a persistent Claude Code integration and receives token-authenticated hook requests; the Stop hook processes assistant transcript content locally.

Mechanism

User-invoked Claude Code MCP and hook installation with bearer-token network calls.

Rationale

Source establishes a user-invoked, first-party Claude Code extension setup with persistent configuration and remote token use. It is not install-time malware, but warrants a warning under the agent extension lifecycle policy.

Evidence

package.jsonbin/kenship.mjsfloor/hook-glance.shfloor/hook-scan.mjsfloor/statusline.sh~/.kenship/credentials~/.kenship/bin~/.claude/settings.json~/.claude/commands/kenship.md~/.zshrc~/.bash_profile~/.bashrc

Network endpoints2

postcard-production-145b.up.railway.app

kenship@0.1.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code