The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and process source code from external, untrusted repositories, creating a potential surface for indirect prompt injection attacks.
Ingestion points: The agent ingests external repository content via search commands like rg (ripgrep) and directory listing tools (ls).
Boundary markers: The skill does not provide specific instructions for the agent to use delimiters or ignore instructions that may be embedded within the files it audits.
Capability inventory: The skill's instructions involve the use of rg, dig, and ls to analyze external data.
Sanitization: There is no mention of sanitizing or filtering input from the target repositories before it is processed by the agent during the audit.
[EXTERNAL_DOWNLOADS]: The skill references various technical standards and libraries from trusted organizations to support the audit workflow.
Evidence: The skill links to official IETF RFCs, MITRE CWE definitions, NIST SP 800-63 documentation, and OWASP cheat sheets.
Evidence: It recommends using Google's official libphonenumber library for country-aware phone number validation.

aegis-email-sms