The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the oh-my-openagent package from a third-party GitHub repository (code-yeongyu/oh-my-openagent) and the lazycodex-ai package from the npm registry using npx during the setup phase.
[REMOTE_CODE_EXECUTION]: The installation process executes code downloaded at runtime from external sources to configure the local environment and set up command aliases.
[COMMAND_EXECUTION]: The skill implements an autonomous coding loop ($ulw-loop) that executes arbitrary shell commands and provides a mechanism for the agent to run scripts directly via the $command syntax. It also includes tools like AST-grep and LSP that interact with the system shell.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it hierarchically ingests context from AGENTS.md and rules files within the project repository.
Ingestion points: Hierarchical AGENTS.md files generated by /init-deep and existing project rules files.
Boundary markers: No explicit delimiters or instructions are used to distinguish project data from agent instructions; content in context files is treated as authoritative guidance.
Capability inventory: The agent has the capability to write to the file system (/init-deep), execute shell commands ($ulw-loop), and perform multi-agent orchestration.
Sanitization: The skill does not perform sanitization or validation of the content processed from project context files.

lazycodex-agent-harness