The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes scripts that ingest and process data from external, untrusted sources which could contain embedded instructions. \n
Ingestion points: The scripts/process.py script reads user-provided email files (.eml) to perform header forensics. The scripts/agent.py script fetches threat details, subjects, and URLs from the Proofpoint Targeted Attack Protection (TAP) API. \n
Capability inventory: Both scripts output content that an AI agent would process to make decisions or summarize security status. scripts/agent.py performs network requests to the Proofpoint API. \n
Sanitization: There is no explicit sanitization for LLM prompts in the script outputs. \n
Boundary markers: No explicit markers are used to wrap external content in the generated reports. \n- [EXTERNAL_DOWNLOADS]: The skill requires standard Python libraries. \n
The scripts/agent.py file and documentation require the requests library for interaction with the Proofpoint API. \n
The scripts/process.py script requires the dnspython library for performing DNS record lookups. \n- [COMMAND_EXECUTION]: The skill provides functional Python scripts that the agent is expected to execute via the command line. \n
scripts/agent.py is used to audit blocked and delivered email threats. \n
scripts/process.py is used to validate domain configuration and analyze email headers.

implementing-proofpoint-email-security-gateway