The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python utility script (scripts/extract_candidates.py) to automate the extraction of pronunciation candidates. The script performs regex-based text analysis on the provided input file.
[PROMPT_INJECTION]: The skill processes untrusted user-supplied text scripts, which serves as a potential surface for indirect prompt injection attacks. 1. Ingestion points: The scripts/extract_candidates.py script and the agent read the content of the user-provided script file. 2. Boundary markers: No explicit boundary markers or 'ignore instructions' wrappers are used for the input text. 3. Capability inventory: The skill executes a local Python script, writes output files (respelled scripts, tables), and invokes external Text-to-Speech tools. 4. Sanitization: The extraction script uses regular expressions to isolate tokens but does not sanitize the input text before the agent reviews it.
[SAFE]: The integration with external Text-to-Speech services (Higgsfield and ElevenLabs) is documented and aligns with the skill's intended purpose.

phoneticize