The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes environment setup commands including git clone, npm install, and npm run dev to prepare the game runtime. It also executes local Python and Bash scripts (scripts/openword_rest.py, scripts/openword_rest.sh) to interact with the game's REST API.
[EXTERNAL_DOWNLOADS]: The skill clones the OpenWord repository from github.com/dinghuanghao/openword.git. This download targets the author's own organization and is consistent with the primary purpose of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill ingests narrative data and world states from the local REST API and uses them to generate natural language updates, creating a surface for potential injection.
Ingestion points: Game state and narrative data retrieved from /api/get_current_game_state and /api/do_action.
Boundary markers: None observed in the instructions for prompt synthesis.
Capability inventory: Shell command execution, package installation, and local network operations.
Sanitization: No evidence of sanitization or filtering of the narrative content before agent processing.

openword-navigator