Mentor Selection Agent

Plugin mentor-selection-agent in BesTLP/MentorSelectionAgent. Commands read from the repo's manifest — 1 month ago.

Skills (12)

• autonomous-mentor-selection

  Use when the user wants mentor selection to run without user confirmation, with AI making requirement and final workbook gate decisions while preserving evidence, review, candidate-shortfall, and workbook validation gates.

• mentor-candidate-selection

  Use to select mentor and program candidates after confirmed requirement analysis, using internal total tables as recall sources and official pages as final evidence.

• mentor-human-review-pack

  Use after selection review passes to prepare one-mentor-at-a-time human review packages from confirmed JSON and evidence records.

• mentor-knowledge-update

  Use when a human points out a mentor-selection error and the system must classify the error, locate the likely knowledge or rule source, and propose a safe update without directly modifying original tables.

• mentor-management

  Use to build non-destructive mentor management workbooks from internal regional total tables, grouped by region, school, and canonical major.

• mentor-requirement-analysis

  Use for the first mandatory step in a mentor-selection case: analyze the student brief, produce a fixed structured requirement table and JSON for each student, define main discipline anchors, interdisciplinary boundaries, risks, and only truly blocking correction items before candidate selection.

• mentor-selection-agent

  Use when running the full internal graduate mentor-selection workflow as an installable Codex plugin. This orchestrates requirement analysis, source evidence extraction, candidate selection, review, human review packaging, workbook generation, and knowledge update proposals.

• mentor-selection-review

  Use to hard-review mentor selection outputs against student needs, interdisciplinary boundaries, source accuracy, knowledge-base rules, and workbook field requirements.

• mentor-selection-workbook-cn

  Use when the user needs a Chinese counselor-facing graduate supervisor screening workbook or 择导 Excel in the latest grouped 13-column template, especially when the deliverable must use 2026 QS rank or NA, bilingual school names in historical-results order with Chinese full name then official English

• mentor-source-evidence

  Use when extracting official university, faculty, program, mentor, tuition, scholarship, deadline, language, RP/SOP, and application-material evidence for mentor selection. Prevents AI summarization of official fields by requiring verbatim or normalized evidence labels.

• mentor-workbook-generation

  Use after human review confirms the selected mentors. This prepares confirmed JSON for final workbook generation and must reuse the existing mentor-selection-workbook-cn skill for the actual 13-column Chinese workbook rules.

• student-management

  Use to build non-destructive student historical-result management copies, grouped by region and canonical major with one workbook per student.

Slash commands (1)

• /mentor-selection-agent:autonomous-mentor-selection

  Run a mentor-selection case without asking the user for gate confirmations; the AI records autonomous decisions and still enforces evidence, review, and workbook checks.

Similar plugins