The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill programmatically generates YARA detection rules from templates and compiles them at runtime using the yara-python library to validate syntax and test matches.
[INDIRECT_PROMPT_INJECTION]: The skill framework processes untrusted binary data from malware samples to identify unique strings and code patterns for signature development.
Ingestion points: Binary data is read from the local filesystem in agent.py and process.py.
Boundary markers: Boundary markers and instructions to ignore embedded content are not present in the ingestion logic.
Capability inventory: The skill can read local files, write generated rules to disk, and perform rule compilation.
Sanitization: Extracted strings are incorporated into rule definitions without validation or sanitization.

performing-yara-rule-development-for-detection