Scan Report
15 /100
teller
Teller integration. Manage data, records, and automate workflows. Use when the user wants to interact with Teller data.
Documentation-only skill for Teller/Membrane integration with all behavior declared; minor supply chain concern from unpinned npm CLI but no direct security impact.
Safe to install
No immediate action required. Consider pinning the CLI version (e.g., @membranehq/[email protected]) for reproducible builds. The Membrane proxy model is a sound security architecture.
Findings 2 items
| Severity | Finding | Location |
|---|---|---|
| Low | Unpinned npm CLI dependency Supply Chain | SKILL.md:24 |
| Info | Unofficial npm package domain Supply Chain | SKILL.md:24 |
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Network | WRITE | WRITE | ✓ Aligned | SKILL.md:38-57 - membrane request command with arbitrary HTTP methods |
| Shell | WRITE | WRITE | ✓ Aligned | SKILL.md:24 - npm install -g @membranehq/cli; SKILL.md:28-34 - membrane login/co… |
| Filesystem | NONE | NONE | — | No file operations declared or observed |
| Environment | NONE | NONE | — | SKILL.md explicitly states 'never ask the user for API keys or tokens' and uses … |
| credential_theft | NONE | NONE | — | SKILL.md:58 - 'Let Membrane handle credentials — never ask the user for API keys… |
2 findings
Medium External URL 外部 URL
https://getmembrane.com SKILL.md:7 Medium External URL 外部 URL
https://teller.com/docs SKILL.md:19 File Tree
1 files · 4.2 KB · 124 lines Markdown 1f · 124L
└─
SKILL.md
Markdown
Dependencies 1 items
| Package | Version | Source | Known Vulns | Notes |
|---|---|---|---|---|
@membranehq/cli | unpinned | npm | No | Global npm install without version constraint |
Security Positives
✓ No hidden or undocumented functionality — all capabilities declared in SKILL.md
✓ Credential management delegated to Membrane (no local secret storage)
✓ Network requests routed through Membrane proxy with server-side auth lifecycle
✓ Best practices documented: prefer pre-built actions over raw API calls
✓ No sensitive file paths (~/.ssh, ~/.aws, .env) accessed
✓ No obfuscation techniques (base64, eval, atob) detected
✓ No credential harvesting or exfiltration behavior
✓ No persistence mechanisms (cron, startup hooks, backdoors)
✓ No reverse shell or C2 communication patterns