The Agent Skills Directory

[Data Access] (SAFE): The skill references local data files (ads/references/benchmarks.md, ads/references/conversion-tracking.md) to establish baselines for its assessment process. This is a legitimate use of local reference data.
[Indirect Prompt Injection] (SAFE): The skill involves evaluating external landing page content from URLs. While this creates a surface for indirect prompt injection, the instructions are focused on passive analysis and scoring, and there is no evidence of instructions that would cause the agent to execute malicious commands based on that content.
[No Code] (SAFE): This skill consists entirely of markdown-based instructions and process descriptions without any associated scripts or executable files.

ads-landing