The Agent Skills Directory

[PROMPT_INJECTION]: The SKILL.md file contains a section titled 'GLOBAL PROTOCOLS' which uses imperative and restrictive language (e.g., 'MANDATORY', 'You must', 'Iron Law of Execution') to define new operational modes and override the agent's default behavioral guidelines and decision-making logic.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external, potentially attacker-controlled infrastructure. This creates a surface where malicious instructions could be embedded in network metadata.
Ingestion points: Data is fetched from external sources via requests in scripts/agent.py (within fingerprint_http and get_ssl_certificate methods) and scripts/process.py (within shodan_host_lookup and ct_log_search).
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands when processing the content of HTTP banners or SSL certificates.
Capability inventory: The skill allows the agent to perform network requests and write findings to local JSON files (threat_infra_report.json).
Sanitization: The retrieved data is formatted as JSON and printed to the console or saved to disk without sanitization of the strings retrieved from the external services.
[EXTERNAL_DOWNLOADS]: The skill scripts facilitate network communication with third-party security services including Shodan, VirusTotal, SecurityTrails, and crt.sh to gather infrastructure intelligence. These operations are transparently documented as part of the skill's primary function.

tracking-threat-actor-infrastructure