Conference Abstract Adaptor

This skill appears to implement the advertised functionality, but it is draft-quality and contains some security gaps you should consider before using it on sensitive data: - The script reads whatever path you pass; it does not check for '../' traversal, canonicalize paths, or restrict input to a safe workspace. A malicious or mistaken path argument could expose sensitive local files. Only run with input files you control, or add path validation (e.g., resolve realpath and ensure it is under an allowed directory). - The script lacks try/except around file operations and will print raw Python stack traces on errors (SKILL.md promised sanitized error messages). Consider adding exception handling to avoid leaking filesystem layout or other details. - There is no sandboxing or execution isolation required by the skill; run it in an isolated environment if abstracts contain sensitive or unpublished content. - The SKILL.md security checklist items (no traversal, sandboxing, sanitized errors) are NOT enforced by the code — treat them as TODOs, not guarantees. If you want to proceed: review or run the included scripts/main.py locally in a sandbox, add input path validation and error handling, and test with non-sensitive files first. If you need, ask the author to update the code to enforce workspace-restricted file access and to catch errors before installing or invoking this skill in production.