The Agent Skills Directory

[SAFE]: The skill is entirely instructional and provides business analytics frameworks for revenue, efficiency, and pipeline measurement. There is no code or automation provided within the skill that could lead to technical security vulnerabilities.
[METADATA_POISONING]: The metadata fields contain links to the original author's and contributor's GitHub profiles and repositories. These are used for attribution purposes and do not contain malicious instructions or redirects to untrusted content.
[INDIRECT_PROMPT_INJECTION]: While the skill instructs the agent to gather context about a user's business (such as CRM tools and pricing models), it does so to provide tailored advice. The skill lack any dangerous capabilities, such as file-writing, network exfiltration, or shell execution, which are necessary to exploit indirect prompt injection surfaces.

gtm-metrics