The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from multiple sources, creating a surface for indirect prompt injection where malicious instructions could be embedded in data files.
Ingestion points: The skill reads track files (SKILL.md, step 1), album READMEs, and previous track files (Cross-Track Referencing section). It also calls a load_override function to ingest a user-supplied lyric-writing-guide.md file.
Boundary markers: The instructions do not define delimiters or specific warnings to the model to treat the ingested file content as data rather than instructions.
Capability inventory: The skill has access to file system tools (Read, Edit, Write, Grep, Glob) and the bitwize-music-mcp tool. It also performs automated tool chaining by invoking suno-engineer upon finalization.
Sanitization: There is no evidence of sanitization, escaping, or validation performed on the ingested text before it is incorporated into the agent's context.

lyric-writer