The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local bash scripts, such as validate-stories.sh and find-stories-by-tag.sh, along with standard command-line tools like cp, git, and ast-grep to manage file operations and code transformations during the graduation process.
[INDIRECT_PROMPT_INJECTION]: The skill acts upon project source code and Storybook files, creating a potential surface for instructions embedded in code to influence the agent.
Ingestion points: The agent and its scripts read .stories.tsx and .tsx files from the project repository.
Boundary markers: No explicit markers or warnings are used to distinguish between data (code) and instructions.
Capability inventory: The skill possesses the ability to read and write files, as well as execute shell commands.
Sanitization: The processed file content is not sanitized for potential prompt injection patterns.
[DYNAMIC_EXECUTION]: The validation script can run tsc and eslint via npx when strict checks are enabled. The use of the --no flag helps mitigate risks by restricting execution to packages already present in the local environment.

sb-ship