The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies and processes local project assets (source code, configuration files, and package manifests) to generate new verification instructions. This constitutes a potential ingestion surface for untrusted data that could influence the generated skill's behavior.
Ingestion points: The skill reads package.json, e2e.config.ts, local navigation structures, and test files within the target project's repository.
Boundary markers: The instructions do not explicitly mandate the use of delimiters or boundary markers for the data ingested from the project files when generating the new skill.
Capability inventory: The skill is capable of writing to the local filesystem (.agents/skills/), creating symlinks, and executing shell commands through the npx e2e utility.
Sanitization: There are no explicit instructions to sanitize or escape project-specific content before it is interpolated into the generated skill's documentation.
[COMMAND_EXECUTION]: The skill lifecycle includes a verification step where the agent executes shell commands to ensure the generated skill works as expected.
Evidence: Instructions require running npx e2e list, npx e2e run, and npx e2e explore to validate the setup and confirm the application is ready for testing.
[EXTERNAL_DOWNLOADS]: The skill references the addition of external extensions to the agent's environment.
Evidence: The skill mentions using npx skills add tester-army/e2e to install necessary testing dependencies. Since the target namespace matches the skill author's name (tester-army), this is identified as a legitimate vendor-provided resource.

create-verification-skill