openakita/skills@tencent-survey

Provide the survey content as plain text using the DSL syntax, where each question specifies its type in brackets like [单选题] or [量表题]. The create_survey tool parses the text and returns a survey ID and hash for the shareable wj.qq.com link.

Use the list_answers tool with the survey ID to retrieve responses. It uses cursor-based pagination, so pass the returned last_answer_id to fetch subsequent pages until a page returns fewer items than per_page.

It requires a token with the wjpt_ prefix, either set via the TENCENT_SURVEY_TOKEN environment variable or obtained through an OAuth device authorization flow. The setup.sh script handles both methods and writes the token into mcporter configuration.

Yes, the create_survey tool supports four scenes: surveys (scene 1), exams (scene 3), assessments (scene 6), and polls (scene 8). Exams support answer keys, scores, and grading rules; assessments use dedicated question type tags.

Updates fail if the question_id was not obtained from get_survey, the survey is not in an editable state, or the text syntax is invalid. The question type tag must immediately follow the title, and the survey must belong to your token's team.