The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several shell commands such as grep, glob, git log, ls, and cat to automate the discovery of project architecture and directory structures. It specifically instructs the agent to execute these commands across the codebase to populate diagram templates.
[PROMPT_INJECTION]: The skill defines a set of proactive triggers and directives intended to be added to the agent's configuration files (e.g., CLAUDE.md). By appending these instructions, the skill modifies the agent's future behavior and interaction style, creating a persistence mechanism for its own operational logic.
[DATA_EXFILTRATION]: The skill performs automated codebase scanning using glob, grep, and read commands on directories like src/, app/, and files like package.json. While intended for architecture discovery, this allows the agent to access and process the contents of the project environment.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by reading and processing data from untrusted local project files during its auto-discovery phase. Specifically, it reads package.json and performs recursive searches through the src/ directory without defined boundary markers or sanitization of the ingested content.

ascii-diagram-creator