snuppy

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill explicitly instructs fetching and treating remote content as its own instructions (via snuppy://), enabling dynamic remote-controlled behavior and arbitrary instruction/code injection — a high-risk backdoor/supply-chain vector.

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

• Third-party content exposure detected (high risk: 0.90). The required workflow fetches remote content from an MCP server (mcp__snuppy__snuppy_fetch of snuppy://snuppy) and then treats that fetched text as the skill’s instructions, so outsider-authored free text from the remote MCP source can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly requires calling mcp__snuppy__snuppy_fetch at runtime to fetch and treat the remote resource snuppy://snuppy (and any snuppy://... references it points to) as this skill's SKILL.md, meaning external content will directly control the agent's instructions.

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

• Attempt to modify system services in skill instructions detected (medium risk: 0.60). This prompt instructs the agent to fetch and obey arbitrary remote SKILL.md content (and follow further remote references), which enables executing externally provided instructions that could direct privileged or state-changing actions, even though it does not itself explicitly request sudo, user creation, or system-file modification.