web-scrape
Pass
Audited by Gen Agent Trust Hub on May 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by ingesting untrusted content from the web and presenting it to the agent. While it aims for clean extraction, it lacks explicit boundary markers to isolate this content from agent instructions.
- Ingestion points: All data-fetching capabilities like fetch_page and extract_text in SKILL.md and schema.yaml.
- Boundary markers: Absent; there are no instructions to wrap scraped content in delimiters or warnings to ignore embedded commands.
- Capability inventory: The skill enables network reads and is associated with the Bash and WebFetch tools in the frontmatter.
- Sanitization: The extract_text behavior removes HTML boilerplate, which provides a basic level of content filtering.
Audit Metadata