The Agent Skills Directory

[COMMAND_EXECUTION]:
The framework relies on executing system commands, primarily git, for version control, state tracking, and source code verification across multiple repositories.
It coordinates project-specific scripts defined in .sdlc/config.md for standard development tasks such as building, testing, linting, and running applications.
The core logic is driven by the sdlc.sh bash script, which performs state management and gate enforcement.
[EXTERNAL_DOWNLOADS]:
The skill includes a maintenance script (adapters/e2e/update.sh) that fetches the e2e package from the NPM registry to vendor its documentation. NPM is a well-known and trusted service registry, making this a safe operation for its intended purpose.
[INDIRECT_PROMPT_INJECTION]:
The toolkit is designed to process untrusted data from external sources, including ticket descriptions, logs, and web pages.
To mitigate injection risks, it contains explicit architectural rules (such as Rule 3 in AGENTS.md) instructing the AI agent to treat all ingested text as data rather than instructions.
[DYNAMIC_EXECUTION]:
The todo manager utilizes Node.js dynamic imports to load its TUI and Web interface components based on user commands.
The e2e testing adapter employs browser initialization scripts (e.g., popup-bridge.ts) to modify runtime environments for testing purposes.

sdlc-kit