Material Grounded Learning Repo
# Material-Grounded Learning Repository **Mode**: Material-Grounded Learning **Inputs**: `learning_materials/material_manifest.md`, extracted materials, `{domain}`, `{duration}`, `{daily_time}`, `{locale}` **Output files**: `00_domain_map.md`, `01_core_concepts/`, `learning_materials/material_index.md`, `learning_materials/material_coverage_map.md`, `learning_materials/material_learning_plan.md`, `progress.md`, `09_sources/claim_ledger.md` --- Build the learning repository from the indexed user materials, not from a generic syllabus. ## Grounding Priority 1. User materials in `learning_materials/raw/`. 2. Extracted content in `learning_materials/extracted/`. 3. Supplemental outside knowledge, only when needed and clearly labeled `Supplemental`. ## Required Outputs 1. `learning_materials/material_index.md` - Index each material by topic and available page, slide, or section. - Mark charts, tables, screenshots, diagrams, and flowcharts. 2. `00_domain_map.md` - Build the knowledge map from material topics and structure. - Include material IDs and locations for each major concept. 3. `01_core_concepts/` - Create concept files only for concepts present in the material or explicitly marked `Supplemental`. 4. `learning_materials/material_coverage_map.md` - Map every learning module back to material IDs and locations. 5. `learning_materials/material_learning_plan.md` - Create a day-by-day plan that follows the material order where useful. 6. `progress.md` - Initialize progress with material coverage, current extraction issues, and next steps. 7. `09_sources/claim_ledger.md` - Log key factual claims derived from materials with Source Type = `Material`, Material ID / Location, Confidence, and Freshness Risk. - Log outside additions with Source Type = `Supplemental`. 8. `START_HERE.md`, `TODAY.md`, and `07_daily_review/day-01.md` - Create the guided Day 1 entry points after the material index is available. - Day 1 must be grounded in the available material index and must not require the learner to open several files before starting. ## Rules - User materials are the primary source. - Do not replace material content with generic domain knowledge. - External additions must be marked `Supplemental`. - Unreadable or unextracted content must remain an unresolved extraction issue. - Do not fabricate page numbers, slide numbers, visual content, citations, or material topics. - If visual content matters but cannot be extracted, mark the related module as `Partially grounded`. - Material-derived claims must be auditable in `09_sources/claim_ledger.md`. - After the material index and repository are created, do not stop after a file summary. Start the guided Day 1 session in chat unless the user explicitly requested scaffold-only mode. - Before or near Day 1, include the Freshness Notice from `templates/{locale}/freshness_notice.md.template` whenever the repository has `09_sources/freshness_log.md`, `09_sources/claims_to_verify.md`, material freshness risk metadata, supplemental claims, high-stakes content, or fast-changing content. - The notice must state the highest freshness risk, recommended review interval, `09_sources/freshness_log.md`, `09_sources/claims_to_verify.md` when claims require verification, and a no-web / no-retrieval disclaimer if current sources were not verified. - The guided session must tell the learner they do not need to open the files first, include one small chat task, an answer template, completion criteria, and note that `progress.md` will be updated after completion. End generated learning modules with the standard Source Notes and freshness footer.
fill the variables
This prompt has 4 variables. Pro fills them into a ready-to-paste prompt for you — no manual find-and-replace.
{domain}{duration}{daily_time}{locale}
Unlock with Pro →when to use it
Community prompt sourced from the open-source GitHub repo vesperchinn/learn-anything-skill (MIT). A "Material Grounded Learning Repo" style prompt — adapt the placeholders and specifics to your task. Imported as-is and not independently retested here, so check the output before relying on it.
tags
roleplaycommunitygeneral
source
vesperchinn/learn-anything-skill · MIT