html-to-image-render
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DATA_EXFILTRATION]: The skill reads the entire content of a local file specified by the user and sends it to a remote API (https://html2image.hanfangyuan.cn/api/html2image). While this is necessary for the rendering service, it creates a risk where sensitive local data could be exfiltrated if the agent is directed to process non-HTML files. Evidence is found in scripts/render-html-to-image.ts where readFile is used followed by a network fetch.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes file content without sanitization or boundary markers, creating a surface for indirect prompt injection if the agent reads malicious instructions from a file.\n
- Ingestion points: Local files read by the scripts/render-html-to-image.ts script via the --input flag.\n
- Boundary markers: None used to isolate file content from instructions.\n
- Capability inventory: File system read/write access and network POST capabilities.\n
- Sanitization: No sanitization or validation of the input content before processing.\n- [EXTERNAL_DOWNLOADS]: The skill's workflow relies on npx -y tsx, which downloads the tsx package from the npm registry at runtime to execute the TypeScript script.
Audit Metadata