The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests data from untrusted local files like CLAUDE.md, .cursorrules, and other agent-specific metadata files to extract intent and instructions. Malicious content in these files could influence the behavior of the current agent during the initialization process.
Ingestion points: Processes content from .claude/, CLAUDE.md, .cursor/, .cursorrules, codex.md, .codex/, .aider*, .continue/, and .copilot/ (SKILL.md).
Boundary markers: No specific delimiters or safety instructions are defined to separate ingested file content from the agent's core instructions.
Capability inventory: The skill performs file system read and write operations to summarize project state and generate new configuration files in the .agent-monad/ directory.
Sanitization: No sanitization or validation of the content extracted from external configuration files is described.

agent-init