The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes several MCP tools to perform text-to-speech operations, specifically mcp__mcp-tts__google_tts, mcp__mcp-tts__openai_tts, mcp__mcp-tts__elevenlabs_tts, and mcp__mcp-tts__say_tts.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it ingests and speaks project-related content which may contain untrusted data.
Ingestion points: Reads text from planning documents, resolved issues, and task summaries within the project files.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the text-to-speech interpolation.
Capability inventory: The skill can perform file system operations (read/write configuration) and call external network-based TTS APIs.
Sanitization: The skill implements a 'Text Transformation' process that removes URLs, code blocks, and long technical identifiers, which mitigates some injection vectors that rely on complex or technical strings.
[SAFE]: Accesses project-specific and global configuration files located at .claude/tts-config.json and ~/.claude/ respectively for persistence of voice assignments and provider status. These operations are standard for a development-focused agent skill.

speak