The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user inputs (product ideas, success metrics, and customer signals) to generate foundational repository files such as AGENTS.md and README.md. These files act as persistent instructions for any AI agent interacting with the repository in the future.
Ingestion points: SKILL.md identifies 'product idea', 'target user or customer signal', and 'repo or app context' as required inputs provided by the user.
Boundary markers: There are no explicit instructions to use delimiters or 'ignore embedded instructions' warnings when incorporating user-provided content into project files.
Capability inventory: The skill performs file system writes to create project scaffolding and executes git commands.
Sanitization: The instructions lack specific requirements for sanitizing or escaping external content before it is written to documentation files.
[COMMAND_EXECUTION]: The skill utilizes shell commands for repository initialization and internal lifecycle tracking.
Evidence: SKILL.md contains the command omh runtime record --skill idea-to-deploy --harness app-delivery-loop --status started to log workflow progress.
Evidence: references/project-bootstrap.md specifies the execution of git init to set up version control and git status --short to verify the state of the repository.

omh-idea-to-deploy