pincushion-mcp@1.11.28
Visual feedback as agent work packets. Stakeholders drop pins on your live app; your AI coding agent reads each pin (selector, screenshot, DOM, thread, acceptance criteria) via MCP and ships the fix.
AI Security Review
scanned 1mo ago · by lpm-firewall-aiSource confirms explicit-user-command agent configuration mutation, which the supplied policy assigns a warning. Consent gates and project-local safeguards provide no basis for classifying this path as malicious install-hook abuse or covert takeover.
Decision evidence
public snapshot- server.js activates editor setup through the explicit setup command.
- setup.mjs requires approval before installing a project runtime and connecting the selected editor.
- lib/setup-core.js writes a Pincushion MCP entry into project editor configuration.
- lib/client-critique-install.js installs project agent skills using exclusive file creation.
- Claude skill installation requires explicit consent; setup calls the client installer after approval.
- lib/setup-critique.js saves project-specific consent and runtime identity for continuation.
- package.json has no automatic install lifecycle hooks or runtime self-dependency.
- Setup preserves global configuration and blocks unsafe paths, conflicting skills, and unsupported editor policies.
- Runtime dependency installation disables npm lifecycle scripts; no concrete malicious behavior was established.
Source & flagged code
12 flaggedA single source file combines environment access, network access, and code or shell execution with blocking evidence.
mobile-wall.mjs#virtual:normalized:round1View on unpkg · L121Findings
Affected versions and remediation
This report applies to pincushion-mcp@1.11.28.
See all versions of pincushion-mcp for other recorded verdicts.
- Review the evidence and your use of pincushion-mcp@1.11.28 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.