The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection.\n
Ingestion points: The agent is instructed to process and refactor legacy codebases, which constitute untrusted external data (SKILL.md).\n
Boundary markers: There are no instructions to use delimiters or ignore embedded instructions within the processed code files.\n
Capability inventory: The skill has extensive shell execution capabilities including git, npm, npx, kubectl, and curl across multiple scripts (SKILL.md, references/compliance-checklist.md).\n
Sanitization: No sanitization or validation of the input source code is specified before processing or executing related commands.\n- [COMMAND_EXECUTION]: The skill provides numerous commands for the agent to execute in the local environment.\n
It includes version control and build commands using git and npm (SKILL.md).\n
It utilizes analysis and linting tools such as complexity-report, eslint, and tsc (references/advanced-techniques.md).\n
It defines deployment and rollback operations using kubectl and curl for interacting with local APIs and feature flag services (references/compliance-checklist.md).\n- [EXTERNAL_DOWNLOADS]: The skill incorporates the use of various external tools and libraries from well-known registries.\n
It uses npx to download and run utilities like jscodeshift, autocannon, artillery, and stryker from the npm registry (SKILL.md, references/compliance-checklist.md).\n
It provides instructions for installing project dependencies via npm, pip, and bundle.

code-refactoring