The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes code snippets that demonstrate how to perform security audits using child_process.execSync to execute grep and git commands. These commands are used to inspect the local filesystem for hardcoded secrets and to verify that sensitive configuration files like .env are correctly handled by version control.
[INDIRECT_PROMPT_INJECTION]: The skill describes an audit process that reads and evaluates content from local source files in the src/ directory, which constitutes an attack surface for indirect prompt injection if those files contain malicious instructions. However, the risk is negligible as the commands use static search patterns.
Ingestion points: Local source files within the src/ directory and .env configuration files.
Boundary markers: None; the audit utility processes file content directly through standard shell tools.
Capability inventory: Shell command execution (grep, git) and authenticated network operations to Apollo.io for connectivity testing.
Sanitization: The implementation uses hardcoded regex patterns and static command arguments, preventing untrusted data from altering the execution logic.

apollo-security-basics