superclawd@0.1.105
Claude Code, supercharged — your team's standards, every session. The superclawd CLI loads your workspace's skills, agents, commands, workflows, and live directives into Claude Code as a plugin.
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe package’s disclosed memory feature has opt-out controls, but its default session hooks process transcript text and tool data automatically, a broader data flow than the README clearly explains. This is a privacy concern warranting a warning rather than a malware block.
Decision evidence
public snapshot- The launcher adds automatic capture hooks for session stop and end when capture is enabled and a repository fingerprint is available.
- The capture helper reads the session transcript, including user and assistant text, tool inputs, and tool results.
- At session end, the helper sends a condensed transcript to a Claude CLI process for memory extraction, then posts generated memories to the SuperClawd API.
- The README describes AI-captured memories and provides `--no-capture` and a Private Memory setting, but does not explain that session transcript content is processed automatically.
- The manifest has no install hook; `prepublishOnly` only builds the package.
- The capture behavior supports the package’s stated cross-session memory feature and has opt-out controls.
- The helper posts extracted memory records to the package’s own API, not a hard-coded unrelated receiver.
Source & flagged code
13 flaggedSource sends credentials or rich application records to a package-controlled external receiver enabled by default.
dist/index.jsView on unpkg · L2Findings
Affected versions and remediation
This report applies to superclawd@0.1.105.
See all versions of superclawd for other recorded verdicts.
- Review the evidence and your use of superclawd@0.1.105 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.