The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the Bash tool to run unit and integration test suites during the quality check phase. This allows for the execution of local scripts or binaries as defined in the project's testing configuration.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing content from various project documents such as Game Design Documents (GDDs), Architecture Decision Records (ADRs), and QA reports to determine if quality standards are met.\n
Ingestion points: The skill reads files from multiple directories including design/gdd/, docs/architecture/, and production/qa/ using Read, Glob, and Grep tools.\n
Boundary markers: There are no specific instructions or delimiters mentioned in the skill definition to isolate data read from these files from the agent's core instructional context, nor are there warnings for subagents to ignore embedded instructions.\n
Capability inventory: The agent possesses powerful capabilities including Bash (command execution), Write (filesystem modification), and Task (delegation to other agents), all of which could be targeted by adversarial instructions embedded in ingested documents.\n
Sanitization: The skill does not explicitly perform sanitization, escaping, or validation of the text content retrieved from project documents before using it to inform the validation verdict or passing it to the 'Director Panel' subagents.

gate-check