The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides a JavaScript implementation for a patch script in references/package-patch.md and instructs the agent to write this code to a local file and execute it to modify the source code of a third-party dependency.
Evidence: The file references/package-patch.md contains the full source for scripts/patch-headless-cookie-domain.mjs. The instructions direct the agent to run node scripts/patch-headless-cookie-domain.mjs /tmp/ig-headless-cookie-patch to perform string replacements on minified JavaScript files (e.g., dist/clientside-DZoD6gVq.mjs) inside the dependency's directory.
[COMMAND_EXECUTION]: The skill requires the agent to execute several shell commands to manage the environment and apply patches.
Evidence: references/package-patch.md contains shell blocks with commands such as pnpm add, pnpm patch, node scripts/patch-headless-cookie-domain.mjs, and pnpm patch-commit.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from merchant configuration and browser cookies, creating a potential surface for indirect injection.
Ingestion points: selected-market cookie values, merchant-provided organizationId, experienceId, and variationId (referenced in SKILL.md).
Boundary markers: The instructions do not define clear delimiters or "ignore embedded instruction" warnings when processing these external inputs.
Capability inventory: The skill can write and execute local scripts, perform shell commands, and conduct network operations via the Intelligems and Shopify Storefront APIs.
Sanitization: While the skill recommends validating cookie values against configured market entries, it lacks specific sanitization for merchant-provided identifiers.

intelligems