The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external project files which could contain malicious instructions designed to manipulate the behavior of the orchestrated subagents.
Ingestion points: The skill reads feature descriptions from arguments and parses project files located in design/gdd/, assets/audio/, production/review-mode.txt, and .claude/docs/technical-preferences.md.
Boundary markers: Absent. The skill does not provide specific delimiters or instructions to subagents to disregard potential commands embedded within the ingested text.
Capability inventory: The orchestrated environment has access to high-privilege tools including Bash, Write, Edit, and Task for spawning additional subagents.
Sanitization: Absent. The skill lacks mechanisms to sanitize, validate, or filter the content of ingested documents before they are interpolated into the prompts for subagents.
[DYNAMIC_EXECUTION]: The skill dynamically determines which subagent to invoke based on content read from a configuration file. The primary engine specialist is identified by reading .claude/docs/technical-preferences.md and then used as a parameter for the Task tool to spawn a subagent.

team-audio