---
name: "sat_act_prep"
description: "Prepares students for the digital SAT and enhanced ACT: test format facts, section-by-section strategies, a practice-test workflow with error logging, an ACT↔SAT score converter, and an 8-week study-plan template. Triggers when a user asks about SAT or ACT prep, score goals, study schedules, practice tests, or test-day strategy."
---

# SAT/ACT Prep

## Purpose
Coach an agent through helping a student prepare for the digital SAT or the enhanced ACT: accurate format knowledge, targeted strategies per section, disciplined practice-test cycles, and a concrete 8-week study plan. Verified facts live in `references/` — never restate format facts from memory without checking them.

## Workflow
1. **Identify the test and the goal.** Ask which test (SAT, ACT, or both), the target test date, current score (or "never tested"), and goal score / goal percentile. If they haven't taken a diagnostic, step 1 is always: take one official full-length practice test under timed conditions (Bluebook for SAT, official ACT practice for ACT).
2. **Set the baseline.** Log the diagnostic section scores and domain breakdown (SAT: Bluebook "My Practice" domain scores; ACT: section scores 1–36). Convert between tests with `bin/score_convert.py` when comparing.
3. **Diagnose, don't generalize.** Sort every missed question by domain/skill (use the error-log template in `references/practice-test-workflow.md`). The top 2 domains by misses become the week's drill targets — content gaps first, pacing gaps second.
4. **Build the plan.** Fill in `assets/eight-week-study-plan.md`: weekly targets, drill assignments (official sources only — Bluebook, Khan Academy Official Digital SAT Prep, ACT official guides), and one timed practice test every 2–3 weeks. Scale to the actual weeks remaining.
5. **Run the practice-test loop.** After each test: same-day score, next-day error review (every miss gets a category: content gap / careless / pacing / misread). Re-drill the top categories before the next test. Never burn all official practice tests early — keep one fresh for the final week.
6. **Taper.** Final week: light review, one final simulation 5–7 days out, sleep and logistics (device charged, Bluebook installed, approved calculator) — no new material.

## Output Contract
- Every format fact cited must match `references/sat-format.md` or `references/act-format.md`. If a fact isn't in references, verify via web search before stating it.
- A study plan must include: weeks remaining, baseline and goal scores, weekly drill targets by domain, practice-test dates, and a taper week. A plan without dates is a draft, not a plan.
- Score conversions are labeled **approximate** — use `bin/score_convert.py`, which prints ranges.

## Operating Rules
1. **Official sources only for drills.** Bluebook practice tests, Khan Academy Official Digital SAT Prep, ACT official guides. Third-party tests are for extra reps, never for score prediction — their scoring doesn't match the official curves/adaptive algorithm.
2. **Adaptive-test tactics differ.** On the SAT, module 1 is the ceiling-setter — accuracy there unlocks the harder (higher-scoring) module 2. On the ACT, there's no adaptivity; steady pacing and answering everything wins (no penalty for wrong answers on either test — never leave a blank).
3. **Content before tricks.** Teach the underlying skill (algebra, punctuation rules, passage reasoning) before the shortcut. Shortcuts fail under time pressure when the concept is shaky.
4. **No confident unverified claims.** Don't invent test dates, percentiles, concordance values, or policy details. Check references or search.
5. **Score realism.** Baseline + consistent drill hours + re-testing is the whole game. Don't promise specific point gains; report what practice scores actually show.