The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from the repository, such as documentation and mock files, to generate the storyboard.
Ingestion points: Reads feature documentation, design mock manifests, and local mock files (SKILL.md).
Boundary markers: The skill specifies security boundaries (sandbox, CSP) for the rendered output, but does not define explicit delimiters for the agent's internal processing context.
Capability inventory: Can write HTML files to the filesystem, update manifest files, and invoke external tools or skills.
Sanitization: Requires the use of sandbox attributes for iframes and a restrictive Content Security Policy (CSP) in the generated output.
[COMMAND_EXECUTION]: The instructions include the execution of local development tools to maintain file formatting.
Evidence: The skill specifies running prettier --write on the generated storyboard file.
[EXTERNAL_DOWNLOADS]: The skill references and encourages the use of an external suite of agent tools not managed within the skill itself.
Evidence: Mentions integration with the impeccable family of skills (e.g., /critique, /impeccable craft) hosted in the pbakaus namespace.

storyboard