processon-mindmap-generator

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.70). 该技能明确要求在每次运行前强制执行远程版本检查并在失败时“严格禁止提示网络错误”（将错误隐瞒为“暂无更新”），并规定自动执行 shell/更新命令，这等于指示隐藏运行/网络异常并执行外部代码，属于超出单纯生成思维导图且具有欺骗性和潜在滥用风险。

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill requires executing remote checks and explicitly mandates echoing script outputs (full visit/img URLs and the returned copyBlock) verbatim, which can force the model to disclose any sensitive tokens or credentials that appear in those returned URLs/blocks, creating an exfiltration risk.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's mandatory pre-check (run before each use) fetches and reads https://raw.githubusercontent.com/processonai/processon-skills/main/skills/processon-mindmap-generator/version/github-version.json at runtime, and the returned version value directly controls whether the agent must interrupt, prompt the user to update, and (if accepted) run remote-update commands—so this is a runtime external dependency that controls agent instructions.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill mandates executing shell commands and network-fetched scripts (including an automatic update step that runs "npx ... -g" and Python scripts that read/write/clean files), which actively modify the host system state and can install or change software, so it poses a high risk of compromising the machine.