skills/mukul975/anthropic-cybersecurity-skills/building-adversary-infrastructure-tracking-system/Gen Agent Trust Hub
building-adversary-infrastructure-tracking-system
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external web-based threat intelligence APIs, which represents an indirect prompt injection surface.
- Ingestion points: The
query_crtsh,query_urlhaus, andquery_threatfoxfunctions inscripts/agent.pyingest JSON data from remote servers (crt.sh,urlhaus-api.abuse.ch, andthreatfox-api.abuse.ch). - Boundary markers: None; the skill does not implement delimiters or safety instructions when processing the external JSON payloads.
- Capability inventory: The script performs network requests (
requests.get/post) and writes the aggregated results to the local filesystem usingjson.dump. - Sanitization: There is no sanitization or validation of the content retrieved from the external APIs before it is included in the final report.
- Contextual Note: This surface is considered safe in this context as the skill targets well-known, reputable security intelligence services and the data ingestion is essential for the tool's primary purpose of threat infrastructure tracking.
Audit Metadata