vision-support

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly asks the user to "填密钥" (input an API key) and documents a command config set-key <name> <key> which requires placing the secret directly on the command line, forcing the agent or user to handle and embed the secret verbatim (high exfiltration risk).

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). This set is mostly official API endpoints and repository links, but it includes a direct raw GitHub shell script URL (curl|bash) used for one‑line installation—downloading and executing a remote shell script is a high‑risk pattern that can be used to distribute malware.

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

• Third-party content exposure detected (medium risk: 0.30). SKILL.md/vision-support’s required runtime path does ingest outsider-authored free text at execution time via scripts/vision.mjs parsing the final CLI argument(s) as prompt (parseArgs), which is then sent as text in provider requests (e.g., callOpenAI/callGoogle/callAnthropic).