schema-markup-generator
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process content from external, untrusted webpages provided by the user. This creates a surface for indirect prompt injection where malicious content on a crawled site could attempt to influence the agent's output, such as injecting malicious scripts into the generated JSON-LD or the resulting GitHub Pull Request.
- Ingestion points: The skill uses Chrome DevTools or curl to fetch content from live URLs as described in Step 2 of SKILL.md.
- Boundary markers: The instructions lack explicit structural delimiters to isolate external content within the agent's prompt context, although they specify using only on-page content.
- Capability inventory: The agent has the ability to crawl URLs, read local HTML files, and open GitHub PRs using the GitHub CLI or API in Step 7 of SKILL.md.
- Sanitization: The skill emphasizes producing valid JSON-LD and flags missing fields, but does not explicitly detail sanitization steps for potentially malicious HTML or JavaScript embedded in the source page content.
Audit Metadata