2026-legal-research-agent
Pass
Audited by Gen Agent Trust Hub on Mar 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of scraping and storing content from external legal websites.\n
- Ingestion points: Content is fetched from external URLs (state legislatures, courts, and legal databases) and stored in JSON files within
src/data/scraped/states/as described inSKILL.mdand processed byscripts/validate-scraped-data.ts.\n - Boundary markers: While the data is stored in structured JSON, the
fullTextandcontentfields do not employ boundary markers or instructions to treat the scraped content strictly as data, potentially allowing embedded instructions to affect the agent's logic.\n - Capability inventory: The skill instructions reference the execution of local scripts for data collection (
scripts/firecrawl/run-p0.ts) and the integration of this data into a downstream eligibility wizard.\n - Sanitization: There is no evidence of sanitization or filtering to remove potential instructional triggers from the scraped text in the provided scripts.
Audit Metadata