The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run various shell commands using pnpm, npm, node, and bun. This includes executing build and test scripts like tsgo, remix test, and a project-specific script node ../../scripts/copy-package-type-only-exports.ts .. These commands are part of the intended monorepo management workflow and target resources within the repository structure.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it incorporates user-supplied package names directly into shell commands and file paths.
Ingestion points: The variable <package-name> is used in SKILL.md to construct directory names, package names in package.json, and arguments for pnpm --filter commands.
Boundary markers: There are no specific boundary markers or instructions to the agent to treat the user input as literal data only or to ignore potential instructions embedded within the name.
Capability inventory: The skill is capable of filesystem writes and executing shell commands with arguments derived from user input.
Sanitization: The instructions do not include any steps for sanitizing, escaping, or validating the package name before it is passed to the shell environment.

add-package