# Independent startup-course grade

VERSION HISTORY: 2026-10-14 EDT | chatgpt_codex / source_reviewer | initial local grade, before identified corrections. DEV_LOG reviews/* coverage parent codex-2026-10-14-startup.

Scope and independence: separate implementation reviewer/student pass; reviewer researched 11 providers and did not write course/tool. Shared startup brief rubric used, not BOTS peer grading. Read 13 lessons, generated tool behavior, all worksheet resources and embedded workshop. Actual desktop/mobile two-persona browser evidence in student-test-2026-10-14.md. No deployed release, customer pilot or marketing QA certified here.

| Criterion | Score | Evidence and deductions |
|---|---:|---|
| Research accuracy | 19/20 | Conservative dated provider claims, restrictions and explicit uncertainty; audited 11 assigned records against official sources, parent refresh covers remainder. One point withheld for unverified acceptance details and remaining ambiguous Vercel cap. |
| Practical usefulness | 18/20 | Workbook, two founder stacks, application/stacking checklist, no-credit budgeting and twelve-role worksheet lead to concrete investigation. Limited numeric real-workload pricing/evaluation worked examples. |
| Course quality | 12/15 | Thirteen SHOW/EXPLAIN/DO/TEST/RESULT sequences, written completion gate and five explanatory quiz items. Many steps remain very short and tests largely reflective rather than measured demonstrations. |
| Currentness | 14/15 | October 14 observed records, accessed dates, explicit closed/historical/redirect cases and 90-day gate. Perplexity retained older observation honestly; availability still requires recheck. |
| Clarity | 9/10 | Source verification separated from availability and acceptance; unknown coverage explicit. Some stale draft/QA wording confuses status. |
| Completeness | 8/10 | Broad 22-provider landscape, situation routes, source ledger, capstone and resources. Live validation, final marketing QA and release evidence not yet completed. |
| Exercises/capstone | 3/5 | Persona exports and progress work; scorecard arithmetic correct. MAJOR invalid score serialization, and scorecard reload loses work. |
| BuildBotsBetter fit | 5/5 | Dedicated identity/folder, port allocations, attribution and source preservation guidance present. |
| **Total** | **88/100** | **Below A; one open MAJOR.** |

Required before re-grade: prevent invalid-score export and retest it; preserve scores or disclose storage scope. Update accurate QA wording. Publish/live/marketing completion should be independently evidenced and reported rather than inferred from this local score. No grade inflation for authorized deadline.

## Re-grade after observed corrections, 2026-10-14 EDT

VERSION HISTORY append: source_reviewer independently retested author fixes. Score six blocks export; score four persists across desktop/mobile reload and exports correctly. Revised status wording inspected. course-extra is confirmed template source embedded in learner index, so no standalone-page requirement applies.

Exercises/capstone increases from 3 to 5: invalid export and data-loss findings resolved. Clarity increases from 9 to 10: current audit scope and remaining unknowns distinguished. Other criteria unchanged, including release/live/marketing completeness deductions and concise course-depth limitations.

**Final local course grade: 91/100 (A), no open BLOCKER/MAJOR from this pass.** Research accuracy 19, practical usefulness 18, course quality 12, currentness 14, clarity 10, completeness 8, exercises/capstone 5, BBB fit 5. This score does not certify public deployment, marketing package approval, application acceptance or actual customer outcomes. Retain original 88 and findings as history.