The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from external threat intelligence feeds which represents a surface for indirect injection if a feed source is compromised or contains malformed data.
Ingestion points: The functions ingest_taxii_feed, ingest_urlhaus_feed, and ingest_feodotracker in scripts/agent.py ingest data from remote TAXII and REST APIs.
Boundary markers: No boundary markers or instructions to ignore embedded commands are present in the processing logic.
Capability inventory: The skill has the capability to write files to the local system (export_stix_bundle) and make authenticated POST requests to internal SIEM and MISP platforms (push_to_splunk_ti and MISP integration snippets).
Sanitization: Input values from the feeds are interpolated directly into STIX patterns (e.g., f"[url:value = '{ioc_value}']") and API payloads without escaping. This could allow characters like single quotes to break the pattern logic in downstream detection systems.
[EXTERNAL_DOWNLOADS]: The skill downloads threat intelligence data from several well-known and reputable security services.
Fetches structured threat data from CISA's Automated Indicator Sharing (AIS) TAXII server.
Downloads malicious URL and IP blocklists from Abuse.ch services (URLhaus and Feodo Tracker).
References community threat pulses from the AlienVault Open Threat Exchange (OTX).
[CREDENTIALS_UNSAFE]: The skill manages API keys and authentication tokens for multiple platforms. All provided code uses placeholders such as YOUR_OTX_API_KEY, your_username, and splunk_token, and correctly suggests using environment variables or configuration for secret management rather than hardcoding credentials.

building-threat-intelligence-feed-integration