dcoon-notebook

Fail

Audited by Snyk on Nov 21, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E004: Prompt injection detected in skill instructions.

  • Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate notebook-access workflow, but its MCP documentation link displays one GitHub repository (rait-09/obsidian-agent-client) while the actual target is a different project (coddingtonbear/obsidian-local-rest-api), a mismatch that could mislead the user about where they are being sent, though the real target appears legitimate and the error looks careless.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (low risk: 0.10). The required workflow searches and reads the user's own Obsidian notebook (local vault via obsidian-mcp, obsidian-cli, or local files), which is first-party content, but notes may contain web-clipped or imported external text that an outsider authored, and the agent reads it without a specific item being chosen in advance.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (medium risk: 0.30). The decisive boundary is that the skill instructs registering and connecting to a loopback MCP endpoint (127.0.0.1) whose tool definitions it then relies on at runtime, but that endpoint is a local community-plugin service with no established first-party provenance, so it is a common/reputable unofficial dependency (medium), while the GitHub documentation link is static and does not count.

MEDIUM W013: Attempt to modify system services in skill instructions.

  • Attempt to modify system services in skill instructions detected (low risk: 0.10). The skill's Install branch instructs the agent to register obsidian-mcp as a remote HTTP MCP server, an agent integration change, but the evidence shows no explicit irreversible operational loss, so it stops at low.

Issues (4)

E004
CRITICAL

Prompt injection detected in skill instructions.

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

W013
MEDIUM

Attempt to modify system services in skill instructions.

Audit Metadata
Risk Level
CRITICAL
Analyzed
Nov 21, 2026, 04:17 PM
Issues
4