The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process script content provided by users or external sources through its Python-based analysis scripts, creating a surface for indirect prompt injection.
Ingestion points: The Python utilities readability_scorer.py and script_timer.py accept file paths to read and analyze script content from the local filesystem.
Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded commands within the ingested script data.
Capability inventory: The analysis scripts are limited to string processing, regex matching, and statistical calculations. They do not utilize network libraries, file system write operations, or dynamic execution functions (e.g., eval, exec).
Sanitization: While the scripts perform regex-based cleaning for analysis purposes, they do not specifically filter for prompt injection patterns.
[METADATA_POISONING]: The skill's documentation contains inconsistencies regarding the names of the provided utility scripts.
Evidence: SKILL.md and Scriptwriting_Skill_Summary.md refer to filenames such as timing_calculator.py, dialogue_formatter.py, and engagement_analyzer.py, whereas the included files are named script_timer.py and readability_scorer.py. This mismatch can lead to confusion about the skill's actual capabilities.

scriptwriting