The Agent Skills Directory

[SAFE]: No security issues or malicious patterns were detected. The skill's implementation aligns with its stated purpose of academic writing assistance.\n- [COMMAND_EXECUTION]: The skill utilizes a local Python script (scripts/style_signature.py) to perform text profiling. The script is well-documented, uses only standard libraries, and performs safe, read-only operations on user-provided files.\n- [DATA_EXPOSURE]: The skill processes local research documents and stores style profiles in a local project directory (.paper-memory/). There is no evidence of access to sensitive credentials, system configuration files, or unauthorized network activity.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text from paper drafts and external exemplars. While this represents a standard injection surface for LLM-based tools, the risk is mitigated by the skill's reliance on quantitative metrics and strict instructions to maintain the technical substance of the draft verbatim.

match-style