The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing the @nsio/nsl package from the NPM registry to facilitate local development URL routing and reverse proxying. It also references the official GitHub gitignore repository for populating project configuration.
[COMMAND_EXECUTION]: The skill utilizes a custom Bash script (scripts/task-state.sh) to manage task states with file locking via flock and text processing with awk and sed. It also orchestrates development processes using tmux for persistence and fuser or ss for process management.
[INDIRECT_PROMPT_INJECTION]: The skill reads and updates task, plan, and changelog files within the repository (docs/task/, docs/plan/), which represents an ingestion surface for potentially untrusted instructions. Ingestion points: Task index and detail files, plan files, and the repository changelog. Boundary markers: The skill does not define specific delimiters for separating project data from instructions. Capability inventory: The skill possesses the capability to execute shell commands, perform file system operations, and initiate network requests for dependency verification. Sanitization: The task-state.sh script includes basic regex validation for the owner identifier to prevent shell injection.
[DYNAMIC_EXECUTION]: The skill is designed to execute and manage project-specific development commands (such as bun run dev, vite, cargo run) at runtime, often wrapped within a reverse proxy or persistent terminal multiplexer based on the project's stack.

pma