registry / @sigloch/graphcode / 0.3.0

@sigloch/graphcode@0.3.0

Claude Code Sidecar Harness — Graph Operations + MCP Tools + Hook System

AI Security Review

scanned 4h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. Explicit CLI setup mutates the current project's Claude Code configuration and installs local hooks. No install-time mutation or exfiltration was found.

Static reason

No blocking static signals were detected.

Trigger

User runs graphcode init or graphcode update.

Impact

Can constrain Claude Code reads and writes in the initialized project.

Mechanism

First-party Claude Code skill and PreToolUse-hook scaffolding.

Rationale

The package has a real agent-control surface when explicitly scaffolded, so it warrants a warning under policy. Source inspection found no unconsented install-time activation, exfiltration, remote execution, or malicious payload.

Evidence

package.jsondist/cli.jsdist/scaffold.js.claude/settings.json.claude/hooks/deny-graph-write.sh.claude/hooks/deny-stale-prose-read.shdist/viewer/host.js.claude/hooks/deny-binary-source.sh.claude/skills.mcp.json.graphcode/

Network endpoints1

127.0.0.1

@sigloch/graphcode@0.3.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code