The Agent Skills Directory

[SAFE]: The skill follows its stated purpose of providing a scoring framework for outbound sales hypotheses. All identified scripts and instructions are consistent with this business logic.
[COMMAND_EXECUTION]: The skill includes a local helper script scripts/score.py used to automate the mathematical scoring logic. The script only uses standard Python libraries (argparse, json, sys), performs no network operations, and limits its file access to the input JSON provided via the command line.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided hypothesis lists. The risk is mitigated by the fact that the processing script only extracts specific quantitative fields and returns a formatted table, without executing instructions contained within the data or performing high-privilege actions based on the content.
[OBFUSCATION]: A thorough review of the instructions and scripts confirmed no use of hidden characters, homoglyphs, or encoded payloads. Ukrainian language triggers use valid Cyrillic characters without Latin substitution.

hypothesis-scoring