The Agent Skills Directory

[SAFE]: The skill is a comprehensive security guide designed to help users secure agent boundaries. It provides specific instructions for trust mapping, webhook security using bearer tokens, and least-privilege account scoping.
[SAFE]: The included implementation example in references/implementation.md follows security best practices, using hmac.compare_digest for constant-time comparison, strict regex-based schema validation, and environment-variable-based secret management.
[SAFE]: The skill explicitly instructs the agent to minimize data transfer, avoid printing secrets to logs or stdout, and ensure that consequential side effects (like financial transactions or data deletion) require human confirmation.
[SAFE]: A detailed negative-test matrix is provided to verify that the agent fails safely when encountering unauthorized requests, malformed payloads, or prompt injection attempts.

lindy-security-basics