The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external marketing data which may be influenced by third-party sources.
Ingestion points: scripts/roas_calculator.py processes data via the --file argument, and scripts/ad_health_scorer.py processes data via the --checks and --multi arguments, which read JSON files from the workspace.
Boundary markers: The instructions do not provide specific delimiters or 'ignore embedded instructions' directives for the agent when interpreting the output generated from these data files.
Capability inventory: The skill's capabilities are limited to the two provided Python scripts, which perform mathematical calculations and print analysis to standard output; they do not have network access or file-writing permissions.
Sanitization: The scripts use standard JSON parsing which prevents technical injection, but there is no natural language sanitization to prevent adversarial content within the metrics files from influencing the agent's subsequent recommendations.

paid-ads