geo-llmstxt
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection during its analysis and generation phases.
- Ingestion points: The skill uses the
WebFetchtool to ingest untrusted content from external websites (SKILL.md, Step 3: Write Descriptions), specifically extracting H1 tags, meta descriptions, and the first few paragraphs. - Boundary markers: There are no instructions in the skill to wrap fetched content in delimiters or to ignore potential instructions embedded within the source HTML.
- Capability inventory: The skill has access to powerful tools including
Bash,Write,Read, andWebFetch, which could be exploited if an attacker successfully injects instructions via a website's metadata. - Sanitization: The skill lacks any content validation or sanitization steps to filter out malicious payloads in the processed text.
Audit Metadata