The Agent Skills Directory

[SAFE]: The skill automates project initialization and state tracking using standard tools (Read, Bash, Write) within the local project directory. Significant actions, such as scaffolding or proceeding to new workflows, require explicit user confirmation via the AskUserQuestion tool.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests content from project files to determine the current development stage.
Ingestion points: The agent reads .claude/docs/technical-preferences.md and design/gdd/game-concept.md, and scans the src/, prototypes/, and backlog/ directories for file existence.
Boundary markers: No specific delimiters or instructions to ignore embedded commands within these files are specified in the instructions.
Capability inventory: The skill is authorized to use Bash (limited to cp and mkdir operations in instructions) and the Write tool to modify project files.
Sanitization: The skill does not explicitly sanitize or validate the content of the read files beyond checking for specific configuration strings or file existence.

start