The Agent Skills Directory

[COMMAND_EXECUTION]: The skill manages project lifecycles by invoking various sub-skills for system setup (cofounder-computer-setup), repository initialization, and testing. It utilizes the mise tool for executing development workflows.
[DATA_EXFILTRATION]: During the deployment phase, the agent is instructed to read the .env file to classify and pass secrets to the app-deploy skill. This behavior is documented as a requirement for the deployment process.
[EXTERNAL_DOWNLOADS]: The startup procedure involves installing system-level development tools such as Homebrew, mise, Podman, and the GitHub CLI via the setup sub-skills.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes external, potentially untrusted data from docs/PRD.md and user-provided tasks.
Ingestion points: Reads project requirements from docs/PRD.md and task lists from docs/TASKS.md.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the processing instructions.
Capability inventory: The skill can perform file system writes, execute system-level setup scripts, and perform cloud deployments.
Sanitization: There are no mentioned mechanisms for sanitizing or validating the content of the project documents before they influence agent behavior.

cofounder-playbook