lead-scoring

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface for indirect prompt injection as it uses the Playwright browser tool to fetch information from third-party sources like LinkedIn, news sites, and company pages (detailed in SKILL.md). Ingestion points: Untrusted external websites accessed via Playwright. Boundary markers: Explicitly provided in references/agent-security.md, which contains instructions to treat all retrieved content as data and never as instructions. Capability inventory: The skill utilizes a browser tool and has write access to the .agents/ directory (SKILL.md). Sanitization: The references/agent-security.md file defines protocols for quoting suspected injection attempts and validating all URLs.
  • [DATA_EXFILTRATION]: The skill implements strict safeguards against exfiltration; references/agent-security.md forbids the agent from storing or echoing credentials and requires redacting sensitive tokens found in inputs.
  • [PROMPT_INJECTION]: No malicious prompt injection patterns were identified. A heuristic detector flag for instruction concealment was analyzed and determined to be a false positive, as the skill's instructions focus on enhancing transparency by distinguishing between verified and unverified account data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 04:07 PM