The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the installation of standard, reputable Python libraries (stix2, taxii2-client, and requests) to handle threat intelligence data and network communications.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill processes external threat intelligence data in JSON format from TAXII feeds and local files (e.g., SKILL.md Step 6 and scripts/agent.py).
Boundary markers: There are no explicit instructions or delimiters used to prevent the agent from interpreting natural language descriptions within STIX objects as commands.
Capability inventory: The skill includes network write operations (publishing to TAXII collections via requests.post) and local file system writes.
Sanitization: While the stix2 library performs schema validation, the use of allow_custom=True in agent.py and api-reference.md increases the surface for schema confusion, where unexpected properties could contain hidden instructions.

implementing-security-information-sharing-with-stix2