Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Walk into an IEP or 504 meeting prepared and effective — the process decoded in plain language, the parent-input statement that gets read, the questions that make goals measurable, and advocacy that stays collaborative. Use when asked prepare me for my child's IEP meeting, what's the difference between an IEP and a 504, how do I disagree with the school's plan, or make sure the accommodations actually happen. Produces the process map, the parent-input statement, the goal-quality checklist, the m
.claude/skills/mohitagw15856-iep-504-meeting-kit/SKILL.md| Model | Eval pass | Runs |
|---|---|---|
| gemini-3.6-flash | 100% | 52 |
| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-05 | ✗→✓ | ▲ Improved | 21% | 0% |
| case-11 | ✗→✓ | ▲ Improved | 72% | 0% |
| case-13 | ✗→✓ | ▲ Improved | 45% | 0% |
| case-17 | ✗→✓ | ▲ Improved | 236% | 0% |
| case-19 | ✗→✓ | ▲ Improved | 50% | 0% |
Special-education meetings seat one amateur at a table of professionals who hold these meetings weekly — and hand that amateur the most important vote. Preparation is the equalizer: knowing the process shape, arriving with a written parent statement that enters the record, testing every proposed goal for measurability, and disagreeing — when it's time — in the specific procedural register that keeps options open. This skill builds that kit. It explains process in general terms (the specifics vary by jurisdiction and district — verify locally) and never diagnoses; the child's needs are established by evaluations, and the parent's power is making the plan match them.
Ask for these if not provided:
Plain-language map with the parent-decision points marked · verify-locally flag]
Drafted: strengths · concerns with dated examples · what we're asking the team to address]
| Proposed item | Measurable/owned? | The question to ask | |---|---|---|
Clarify: "How will we measure that, and from what baseline?" · Implementation: "Who ensures each teacher knows, and who checks?" · Disagree: "the noted-disagreement line + options ask]" · Pace: "I'd like to review this at home before signing — what's the timeline for that?"
The recap email template · the progress-report calendar · the plan-isn't-working trigger and next step]
> Special-education processes, timelines, and parent rights vary by jurisdiction and district — verify the local specifics (parent advocacy organizations and the district's own procedural-safeguards document are the standard sources). This kit organizes advocacy; it does not diagnose or give legal advice.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 35,991 | 28,083 | -22% | 1 | 1 | 0% | 6,019 | 5,727 | -5% | 0 | 0 | — |
case-02 | fail→fail | 26,228 | 22,120 | -16% | 1 | 1 | 0% | 4,188 | 5,161 | +23% | 0 | 0 | — |
case-03 | fail→fail | 33,761 | 22,003 | -35% | 1 | 1 | 0% | 5,912 | 5,618 | -5% | 0 | 0 | — |
case-04 | pass→pass | 10,273 | 7,135 | -31% | 1 | 1 | 0% | 1,765 | 2,613 | +48% | 0 | 0 | — |
case-05 | fail→pass | 19,017 | 16,614 | -13% | 1 | 1 | 0% | 3,554 | 4,307 | +21% | 0 | 0 | — |
case-06 | fail→fail | 22,745 | 20,106 | -12% | 1 | 1 | 0% | 3,559 | 4,405 | +24% | 0 | 0 | — |
case-07 | pass→pass | 12,171 | 20,382 | +67% | 1 | 1 | 0% | 2,214 | 4,620 | +109% | 0 | 0 | — |
case-08 | pass→pass | 14,463 | 14,783 | +2% | 1 | 1 | 0% | 2,353 | 3,918 | +67% | 0 | 0 | — |
case-09 | pass→pass | 13,737 | 19,275 | +40% | 1 | 1 | 0% | 2,444 | 4,199 | +72% | 0 | 0 | — |
case-10 | fail→fail | 13,297 | 16,093 | +21% | 1 | 1 | 0% | 2,015 | 3,771 | +87% | 0 | 0 | — |
case-11 | fail→pass | 10,818 | 10,565 | -2% | 1 | 1 | 0% | 1,761 | 3,029 | +72% | 0 | 0 | — |
case-12 | pass→pass | 12,358 | 10,781 | -13% | 1 | 1 | 0% | 2,057 | 3,245 | +58% | 0 | 0 | — |
case-13 | fail→pass | 12,121 | 10,677 | -12% | 1 | 1 | 0% | 2,119 | 3,074 | +45% | 0 | 0 | — |
case-14 | pass→pass | 10,580 | 10,806 | +2% | 1 | 1 | 0% | 1,863 | 3,220 | +73% | 0 | 0 | — |
case-15 | pass→pass | 11,582 | 18,295 | +58% | 1 | 1 | 0% | 2,015 | 4,216 | +109% | 0 | 0 | — |
case-16 | pass→pass | 12,811 | 11,702 | -9% | 1 | 1 | 0% | 2,333 | 3,414 | +46% | 0 | 0 | — |
case-17 | fail→pass | 6,727 | 14,181 | +111% | 1 | 1 | 0% | 1,091 | 3,671 | +236% | 0 | 0 | — |
case-18 | pass→pass | 11,754 | 15,378 | +31% | 1 | 1 | 0% | 1,946 | 3,841 | +97% | 0 | 0 | — |
case-19 | fail→pass | 13,082 | 11,425 | -13% | 1 | 1 | 0% | 2,082 | 3,129 | +50% | 0 | 0 | — |
case-20 | pass→pass | 11,819 | 8,879 | -25% | 1 | 1 | 0% | 1,914 | 2,803 | +46% | 0 | 0 | — |
case-21 | fail→pass | 10,503 | 22,244 | +112% | 1 | 1 | 0% | 2,180 | 5,098 | +134% | 0 | 0 | — |
case-22 | pass→pass | 9,698 | 12,019 | +24% | 1 | 1 | 0% | 1,523 | 3,258 | +114% | 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 +27 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.