The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process external performance data to inform its creative generation tasks.
Ingestion points: The skill processes external data provided via CSV, pasted text, or API outputs (e.g., from google-ads.js) as specified in the "Imported: Iterating from Performance Data" section of SKILL.md.
Boundary markers: No explicit boundary markers or delimiters are used to isolate ingested data from the core instructions, and there are no specific warnings for the agent to ignore embedded instructions within that data.
Capability inventory: The skill possesses the capability to generate text-based ad variations and suggests executable CLI command templates for platform management.
Sanitization: The instructions do not define any sanitization, validation, or filtering logic for the ingested performance data before it is processed by the agent.
[COMMAND_EXECUTION]: The skill provides functional shell command templates for interacting with advertising platform command-line tools.
Evidence: SKILL.md includes a bash workflow example: node tools/clis/google-ads.js reports get --type ad_performance --date-range last_30_days.
[EXTERNAL_DOWNLOADS]: The skill provides documentation and instructions for users to set up external generative AI tools and repositories.
Evidence: references/generative-tools.md includes setup instructions for third-party tools, such as npx create-video@latest and git clone https://github.com/jamiepine/voicebox.git followed by make setup.

ad-creative-v2