The Agent Skills Directory

[SAFE]: The skill provides a rigid mathematical framework for risk assessment aligned with financial standards like Basel III and FATF. It does not request access to sensitive data, credentials, or network resources.- [NO_CODE]: The skill is composed entirely of informational markdown and documentation. It contains no scripts, binaries, or dynamic code generation.- [INDIRECT_PROMPT_INJECTION]: The skill processes external customer profile data (geographic residence, occupation, product types, and onboarding channels) to generate risk scores. This represents a potential surface for indirect prompt injection via untrusted inputs. Evidence Chain: 1) Ingestion Points: User-provided customer profiles in SKILL.md factors. 2) Boundary Markers: None specified. 3) Capability Inventory: None (no subprocess calls, file writes, or network operations). 4) Sanitization: None mentioned. Risk is mitigated by the deterministic formula and a mandatory Human-In-The-Loop (HITL) checkpoint.

risk-assessment