The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The reference file references/ci-integration.md includes an integration example for the well-known service Codecov, which involves fetching and executing a script from https://codecov.io/bash. This is a standard and recognized integration pattern for this service.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a potential attack surface by reading and processing project files. 1. Ingestion points: Project source code and test files read via Read, Grep, and Glob tools. 2. Boundary markers: None explicitly defined in the skill instructions. 3. Capability inventory: Execution through a restricted Bash tool and file modifications via Write and Edit tools. 4. Sanitization: None implemented within the skill; command execution is governed by platform-level allowed-tools restrictions that limit execution to specific test-related command prefixes.

test-automation-expert