tang poetry challenge

唐诗三百首对诗挑战游戏，提供诗句上句，用户需要对出下句。包含挑战结束评分和错题分析功能。

This skill appears coherent and safe: it uses a local JSON of poems and two Python scripts to run a Tang-poetry quiz, with no network calls or credential access. Before installing or running, review and/or run the scripts in a sandboxed environment because running Python code executes locally and can perform arbitrary actions if modified. Note two functional issues in the code (not security problems): 1) calculate_score and get_incorrect_questions call check_answer repeatedly, and check_answer appends to user_answers/correct_answers on each call — this can lead to duplicated/miscounted answers and incorrect scoring/analysis; 2) feedback keyword extraction is a simple heuristic (may not work well with Chinese without tokenization). If you plan to use this skill in production, consider fixing the check_answer side-effect (avoid mutating answer lists when only checking) and validating behavior in a test run. If you have low trust in the source, run the code in an isolated environment (container or VM) or inspect the files locally first.

Type: OpenClaw Skill Name: tang-poetry-challenge Version: 1.0.0 The skill bundle implements a legitimate Tang poetry challenge game. The Python scripts (scripts/challenge.py and scripts/feedback.py) handle game logic, scoring, and educational feedback using a local JSON dataset (references/tang_poetry.json). No evidence of data exfiltration, malicious execution, or prompt injection was found.