The Agent Skills Directory

[SAFE]: The skill is restricted to a read-only toolset consisting of Read, Glob, and Grep, which prevents the agent from modifying the filesystem or exfiltrating data via the network.
[EXTERNAL_DOWNLOADS]: Includes informational links to well-known legal organizations such as the American Bar Association (ABA), the Federal Trade Commission (FTC), and CommonPaper; these references are for guidance purposes and are considered safe.
[SAFE]: No prompt injection patterns, role-play instructions, or attempts to override system safety guidelines were detected in the skill's logic.
[SAFE]: The skill possesses an indirect prompt injection surface as it ingests untrusted contract data (Ingestion points: Read tool for contract files; Boundary markers: absent; Capability inventory: Read, Glob, Grep; Sanitization: absent). However, the risk is negligible due to the absence of write or network capabilities.

risk-analysis