youtube-to-blog-post
Pass
Audited by Gen Agent Trust Hub on Feb 22, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONNO_CODE
Full Analysis
- PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection because it ingests untrusted data from external sources.
- Ingestion points: YouTube video titles and descriptions are retrieved and processed based on user-provided URLs (Evidence: README.md).
- Boundary markers: Documentation mentions
sanitize_text_for_yamlfor front matter safety, but there are no defined delimiters or instructions to ignore embedded commands within the main article body content (Evidence: SEO_UPGRADE_SUMMARY.md). - Capability inventory: The skill has the capability to write files to the local directory (
source/_posts/) and suggests the execution of shell commands for site deployment (hexo cl && hexo g && hexo d) (Evidence: README.md). - Sanitization: While the skill filters YAML special characters to prevent parsing errors, it does not describe sanitization techniques to prevent the LLM from executing instructions embedded in video metadata (Evidence: SEO_UPGRADE_SUMMARY.md).
- DATA_EXFILTRATION (LOW): The skill documentation and configuration templates reference a non-whitelisted domain (
869hr.uk) for image CDN services and related links. While this appears to be part of the author's infrastructure, network operations to non-whitelisted domains are classified as a low-level concern. - NO_CODE (SAFE): No executable script files (e.g., .py, .js) were provided in the analyzed content, although they are referenced in the documentation. Analysis is based on the provided metadata and documented behavior.
Audit Metadata