The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to analyze and process untrusted external data such as brand strategies, datasets, and reference materials. This creates a surface for indirect prompt injection where malicious instructions embedded in these materials could attempt to influence the agent's behavior.
Ingestion points: Untrusted data enters the agent context through the analysis of uploaded videos, brand datasets, and project inputs as defined in SKILL.md.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within processed data.
Capability inventory: The agent can perform file-system operations (writing project files) via the scripts/ogilvideo-brief.mjs helper script.
Sanitization: There is no documented validation or sanitization process for the input data before it is interpolated into the agent's planning and generation logic.
[COMMAND_EXECUTION]: The skill includes automation scripts for production management. scripts/ogilvideo-brief.mjs creates a suite of markdown files for project scaffolding based on user-provided parameters.
[COMMAND_EXECUTION]: scripts/install-ogilvideo-hooks.sh enables persistence within a development workflow by installing a Git pre-push hook. This hook automatically executes a local validation script (scripts/check-ogilvideo.mjs) to ensure project requirements are met before code is pushed.

ogilvideo