中文 EN

Scan Report

Scan New Files

⚠ Low Risk — Risk Score 15/100
Last scan:1 day ago Rescan
15 /100
plaid
Plaid integration for managing banking data, transactions, and workflows via Membrane CLI
Single-file Plaid integration skill using Membrane CLI; all functionality is documented with no hidden behavior, though npm install lacks version pinning.
Skill Nameplaid
Duration24.9s
Enginepi
✓
Safe to install
Pin the CLI version in the install command (e.g., `@membranehq/[email protected]`) to prevent unexpected updates.

Findings 1 items

Severity Finding Location
Low
Unpinned npm package version Supply Chain
The command `npm install -g @membranehq/cli` installs the latest version of the Membrane CLI without version pinning. This could allow a malicious actor to publish a new version with compromised code that gets installed on update.
npm install -g @membranehq/cli
→ Pin to a specific version: `npm install -g @membranehq/[email protected]` or use a hash-based verification.
 SKILL.md:26
ResourceDeclaredInferredStatusEvidence
Filesystem NONE NONE — No file operations described or performed
Network READ READ ✓ Aligned Membrane proxy requests to Plaid API; documented in SKILL.md
Shell WRITE WRITE ✓ Aligned npm install and membrane CLI commands; declared in SKILL.md
Environment NONE NONE — No environment variable access documented or observed
Skill Invoke NONE NONE — No cross-skill invocation
Clipboard NONE NONE — No clipboard operations
Browser NONE NONE — OAuth flow uses external browser, not automated browser control
Database NONE NONE — No database access
2 findings
🔗
Medium External URL 外部 URL
https://getmembrane.com
SKILL.md:7
🔗
Medium External URL 外部 URL
https://plaid.com/docs/
SKILL.md:19

File Tree

1 files · 4.4 KB · 130 lines
Markdown 1f · 130L
└─ 📝 SKILL.md Markdown 130L · 4.4 KB

Dependencies 1 items

PackageVersionSourceKnown VulnsNotes
@membranehq/cli * npm No Version not pinned — installs latest on npm install

Security Positives

✓ All capabilities explicitly documented in SKILL.md — no hidden behavior
✓ Credential management delegated to Membrane's server-side auth lifecycle — no local secrets stored
✓ No credential harvesting, data exfiltration, or obfuscation observed
✓ No sensitive file paths (~/.ssh, ~/.aws, .env) accessed
✓ Uses pre-built Membrane actions rather than raw API calls, reducing attack surface
✓ MIT license and public repository listed for verification