The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script (scripts/campaign_tracker.py) to process JSON campaign data and print reports. The script uses standard libraries and performs no dangerous operations. It handles user input by reading a JSON file path from the command line, which is a common and safe pattern for this type of tool.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were found. The skill does not perform any network operations or download external scripts.
[DATA_EXFILTRATION]: No network access or data exfiltration attempts were detected. The skill operates locally on provided campaign data.
[PROMPT_INJECTION]: The instructions are focused on routing marketing requests and orchestrating campaigns. No attempts to override system safety guidelines or extract system prompts were found.
[OBFUSCATION]: No obfuscated code, hidden URLs, or suspicious character encodings were identified.
[PRIVILEGE_ESCALATION]: The skill does not attempt to escalate privileges or perform administrative tasks.
[PERSISTENCE]: The skill does not attempt to establish persistence on the host system.
[METADATA_POISONING]: The metadata is consistent with the skill's stated purpose and contains no malicious instructions.
[INDIRECT_PROMPT_INJECTION]: The skill processes marketing campaign data (JSON). While this is an ingestion point for external data, the capabilities are limited to status reporting and routing, posing no significant security risk. The script only reads and analyzes the structure of the JSON without executing content within it.

marketing-ops