The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from a URL or local file to drive code generation and PR creation.
Ingestion points: The skill accepts a <handoff-url | path-to-bundle.json> as its primary argument, which is then processed by the design-import skill and subsequent orchestration logic.
Boundary markers: The templates for the PR body and the subagent prompts for test generation do not include explicit delimiters or instructions to ignore embedded malicious prompts within the design data.
Capability inventory: The skill uses the Bash tool to execute git and gh commands, uses Write to modify local files, and uses the Agent tool to trigger subagents for test generation.
Sanitization: There is no evidence of sanitization or strict schema validation for the external bundle data before it is interpolated into agent prompts or PR descriptions.
[COMMAND_EXECUTION]: The skill uses the Bash tool to perform repository management and open pull requests.
The skill executes git status, git branch, and gh pr create to automate the shipping workflow.
These are standard operations for a developer tool and align with the skill's stated purpose.

design-ship