The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted binary data from PCAP files, which could contain embedded instructions designed to influence the agent's behavior during analysis.
Ingestion points: scripts/agent.py and SKILL.md use rdpcap and dpkt to parse network capture files.
Boundary markers: Absent. The skill does not provide instructions to the agent to treat data from the PCAP as non-authoritative or to use specific delimiters.
Capability inventory: The skill has network capabilities via requests and provides instructions for file system interaction such as Suricata rule generation.
Sanitization: Absent. The scripts decode binary payloads to strings without sanitizing them for potential prompt injection patterns.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of external Python libraries and tools.
Evidence: Instructions mention installing scapy, dpkt, and cobalt-strike-parser via pip without version pinning. These are well-known libraries in the security community.

analyzing-command-and-control-communication