The Agent Skills Directory

[SAFE]: The skill implements a prescriptive project management workflow that inspects repository state and requires user confirmation for significant actions, such as advancing to a new production stage.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to run project-local test suites and to update the production/stage.txt file. These operations are restricted to the local environment and are triggered as part of the formal validation process.
[INDIRECT_PROMPT_INJECTION]: The skill processes various project-local documentation (e.g., GDDs, ADRs, playtest reports), which represents a data ingestion surface.
Ingestion points: Reads project files via Read, Glob, and Grep across directories like design/, docs/, and production/.
Boundary markers: The skill does not explicitly utilize boundary markers or safety instructions when processing the content of these files.
Capability inventory: The agent has access to Bash for command execution, Write for file modification, and Task for sub-agent creation, though these are used within the scope of the gated workflow.
Sanitization: No specific sanitization or filtering is applied to the ingested text before it is analyzed by the model.

gate-check