Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Draft SMART IEP goals, accommodations, and present-levels statements that are measurable and compliant in spirit. Use when asked to write an IEP goal, draft special-education goals, list accommodations, or write a present-levels (PLAAFP) statement. Produces measurable annual goals with baselines, criteria, and measurement methods, plus matched accommodations. A drafting aid for educators — not legal advice; the IEP team and local requirements govern.
.claude/skills/mohitagw15856-iep-goal-support/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -1% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 26% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 0% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 18% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -12% | 0% |
IEP goals only help a student if they're measurable: a baseline, a target, a timeframe, and how progress is checked. This skill drafts SMART goals and matched accommodations educators can bring to the team. This is a drafting aid, not legal advice — the IEP team, the student's data, and local/IDEA requirements govern the final document.
Given a student profile and area of need, draft full goals anyway, using clearly-labelled illustrative baselines (replace with the student's real data). Never invent specific diagnoses; work from the need described. Always keep the disclaimer.
Ask for (if not already provided):
A concise, strengths-first paragraph: what the student can currently do, the baseline data, and how the need affects access to the general curriculum.
For each: "By [date], given [condition], [student] will [observable behaviour] to [criterion], as measured by [method] across [n] occasions."
2–4 steps that ladder up to the annual goal.
Matched to the need (e.g. extended time, text-to-speech, chunked tasks, movement breaks) — distinguish accommodations (access) from modifications (changed expectations).
What data is collected, how often, and what counts as on-track vs needs-revision.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 16,254 | 13,004 | -20% | 1 | 1 | 0% | 3,071 | 3,053 | -1% | 0 | 0 | — |
case-02 | fail→pass | 14,762 | 15,312 | +4% | 1 | 1 | 0% | 2,519 | 3,174 | +26% | 0 | 0 | — |
case-03 | fail→pass | 16,399 | 13,794 | -16% | 1 | 1 | 0% | 3,042 | 3,050 | +0% | 0 | 0 | — |
case-04 | pass→pass | 14,332 | 13,702 | -4% | 1 | 1 | 0% | 2,793 | 3,218 | +15% | 0 | 0 | — |
case-05 | fail→fail | 6,955 | 5,134 | -26% | 1 | 1 | 0% | 1,300 | 1,550 | +19% | 0 | 0 | — |
case-06 | pass→pass | 16,185 | 12,024 | -26% | 1 | 1 | 0% | 3,137 | 3,264 | +4% | 0 | 0 | — |
case-07 | fail→pass | 15,690 | 14,089 | -10% | 1 | 1 | 0% | 2,736 | 3,225 | +18% | 0 | 0 | — |
case-08 | fail→pass | 19,547 | 13,093 | -33% | 1 | 1 | 0% | 3,249 | 2,848 | -12% | 0 | 0 | — |
case-09 | pass→pass | 12,824 | 12,426 | -3% | 1 | 1 | 0% | 2,436 | 2,860 | +17% | 0 | 0 | — |
case-10 | fail→pass | 14,240 | 13,121 | -8% | 1 | 1 | 0% | 2,642 | 2,952 | +12% | 0 | 0 | — |
case-11 | fail→pass | 14,297 | 14,344 | +0% | 1 | 1 | 0% | 2,419 | 3,027 | +25% | 0 | 0 | — |
case-12 | fail→pass | 14,372 | 14,569 | +1% | 1 | 1 | 0% | 2,499 | 3,027 | +21% | 0 | 0 | — |
case-13 | fail→pass | 16,491 | 15,226 | -8% | 1 | 1 | 0% | 2,794 | 3,195 | +14% | 0 | 0 | — |
case-14 | fail→pass | 14,938 | 15,588 | +4% | 1 | 1 | 0% | 2,478 | 3,123 | +26% | 0 | 0 | — |
case-15 | fail→pass | 9,278 | 13,392 | +44% | 1 | 1 | 0% | 1,773 | 2,818 | +59% | 0 | 0 | — |
case-16 | fail→pass | 12,232 | 12,816 | +5% | 1 | 1 | 0% | 2,104 | 2,929 | +39% | 0 | 0 | — |
case-17 | fail→pass | 12,405 | 12,475 | +1% | 1 | 1 | 0% | 2,403 | 2,815 | +17% | 0 | 0 | — |
case-18 | fail→pass | 15,863 | 12,173 | -23% | 1 | 1 | 0% | 2,835 | 2,826 | -0% | 0 | 0 | — |
case-19 | fail→pass | 13,237 | 15,751 | +19% | 1 | 1 | 0% | 2,266 | 3,432 | +51% | 0 | 0 | — |
case-20 | fail→pass | 11,783 | 9,771 | -17% | 1 | 1 | 0% | 2,244 | 2,360 | +5% | 0 | 0 | — |
case-21 | fail→pass | 11,200 | 10,642 | -5% | 1 | 1 | 0% | 2,107 | 2,378 | +13% | 0 | 0 | — |
case-22 | fail→pass | 15,816 | 12,880 | -19% | 1 | 1 | 0% | 2,713 | 2,835 | +4% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +82 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.