Comprehensive Plan

A thin orchestration layer over the planfi MCP (https://ai.planfi.app/mcp/free). All math + the 150-year Shiller market dataset live server-side. This skill only gathers inputs and calls the tools — it does not compute anything locally, carries no business logic, math, or defaults, and is read-only (it never changes the user's data). The server is the source of truth.

This is the "one document" superset: it chains four engine sub-analyses — retirement/FIRE, 529 college funding, estate-tax exposure, and insurance/protection — into a single cohesive plan where every dollar is engine-computed. For a FIRE-only deep dive (savings/retirement-age/spend trade-offs, goal-solving, scenario comparison, "what if I change X" forks), use the financial-forecast skill instead; come here when the user wants the broad, multi-section plan.

Step 0 — Make sure the planfi tools are connected

This skill uses these tools (may be namespaced, e.g. mcp__planfi__assemble_comprehensive_plan): assemble_comprehensive_plan (the orchestrator), plus generate_financial_plan, what_if_plan, generate_financial_insights, generate_action_plan, analyze_estate_exposure, analyze_529_optimization, analyze_insurance_needs, analyze_education_credits, and run_backtesting for follow-on deep dives. Use whichever name your environment exposes (bare or mcp__planfi__-prefixed); below they are written bare.

If they're NOT available, tell the user to connect the MCP, then continue:

claude mcp add --transport http planfi https://ai.planfi.app/mcp/free

Try free, then add your key. The command above adds the free connector — https://ai.planfi.app/mcp/free (no key needed). Once you create an API key, add a new connector with the MCP url — https://ai.planfi.app/mcp — and authorize it with your key.

(On claude.ai: add a custom connector pointing at https://ai.planfi.app/mcp/free.)

Access — free for personal use. The planfi MCP is free to try (a small monthly allowance, no key needed). Heavy automated abuse forced us to add limits — but it stays free for personal use: email kam@rateapi.dev and we'll send you a free API key, no charge. (Companies and commercial use have paid plans.) To use a key, pass it as an Authorization: Bearer pft_… header in your MCP client config.

SKILL ROUTING

"build me a financial plan" / "give me a complete / comprehensive / full plan" / "I want one plan covering retirement, college, estate, and insurance" / "do a full financial review or checkup" / "look at my whole financial picture" / "am I on track overall — retirement, kids' college, insurance, everything" → assemble_comprehensive_plan

Always CALL assemble_comprehensive_plan for these — do not answer from general knowledge or quote rules of thumb from memory (no "save 25x expenses", no "$X per kid for college", no "10x income in life insurance", no estate-exemption figures from memory). When the user gives the numbers, run the tool and lead with its real output, THEN explain. This single call fans the household model out to all four sub-engines and returns one cohesive deliverable: cash-flow / retirement Monte Carlo, 529 funding status, estate-tax exposure, and protection gaps — every figure engine-computed. A bare "build me a financial plan" is exactly this tool's job; reach for it first, then enrich.

Gather inputs per Step 1 (only ages/salaries + portfolio are strictly required — everything else is defaulted server-side and read back), then make the primary call per Step 2. Cross-link: for a FIRE-only deep dive or scenario forks, hand off to the financial-forecast skill; to drill into one section of the assembled plan, use the enrichment routes below (chained via { plan_id }).

Step 1 — Gather inputs (only two areas are strictly required)

REQUIRED — ask if not volunteered:

1. Each earner's age + annual salary (household, 1–4 earners).
2. Stock / investment portfolio: current_value + monthly_contribution.

Plan-shaping inputs you may not have (retirement age, desired spend, SWR, returns, inflation, children[] for 529, educationAccount, real_estate[] + mortgages for estate/protection, existing insurance coverage, filing status / state for estate tax): gather them if the conversation makes it natural, but you don't have to chase them down. Anything omitted is defaulted server-side and reported back so the user can correct it (see Step 4) — the server is the source of truth for what was assumed; don't track defaults yourself.

Engine facts to bake in: all dollars are today's (real) dollars; all decimals are fractions (7% → 0.07, 5% → 0.05); tax brackets/limits are approximate ~2026 values (noted in each tool's disclosures).

Optional but recommended — mint a plan_id + share_url: call generate_financial_plan with the household model, CAPTURE the returned plan_id and share_url. Pass { plan_id } to the follow-on deep dives instead of re-sending the model. assemble_comprehensive_plan also accepts the household model directly and runs cold.

Step 2 — Assemble the comprehensive plan (PRIMARY CALL)

Call assemble_comprehensive_plan with the gathered household model (or { plan_id }):

assemble_comprehensive_plan({
  earners: [{ age: 40, annual_salary: 150000, retirement_age: 65 }],
  stocks: { current_value: 300000, monthly_contribution: 3000 },
  desired_annual_spend: 60000,
  children: [{ current_age: 5 }],
  filing_status: "married"
})

It returns one envelope with four nested sections — surface each:

• forecast — FIRE age/year, projected net worth at retirement, Monte Carlo failure / success rate (this is the honest risk check; if it's still Calculating…/undefined, say so — never report a fabricated 0).
• education — total college cost, funded-from-529, funding shortfall, funded %.
• estate — projected estate, applicable exemption, federal + state estate tax, effective rate.
• protection — total life-coverage gap, recommended vs existing coverage, per-earner gaps, and any monthly disability shortfall.
• headline — the engine-composed one-liner across all four sections; lead with it.

Step 3 — Enrich each section on demand (chained via { plan_id })

Only when the user wants to drill into one section — each of these is its own routed intent:

"what if we change one thing — contributions, retirement age, salary, spend — against the saved plan" / "scale back our 401ks next year, what happens to the full plan" → what_if_plan

Always CALL what_if_plan for a single-change scenario against the saved plan — never re-send the household and never narrate a scenario outcome the tool did not compute. Pass plan_id + a small changes array of typed ops; unchanged fields are inherited automatically and the result includes a computed scenario_summary (baseline vs scenario, delta, FIRE ages) to quote verbatim. For the FULL document under the new assumptions, chain the returned scenario plan_id into assemble_comprehensive_plan.

"what's my chance of running out of money" / "run the Monte Carlo" / "stress-test my retirement against history" / "would my portfolio have survived 1929 or the 1970s" / "how safe is a 4% withdrawal for me" → run_backtesting

Always CALL run_backtesting for these — do not answer from general knowledge or quote a rule-of-thumb success rate ("the 4% rule works 95% of the time") from memory. When the user gives the numbers, run it and lead with its real output (the Shiller 1871-present failure rate plus worst/best portfolio longevity), THEN explain. Takes raw portfolio_value / annual_spend / current_age from the plan summary (or chain { plan_id }). This is the honest risk check behind the plan's forecast section — if the rate is still Calculating…/undefined, say so; never substitute 0. Cross-link: for goal-solving around the failure rate (save more / retire later / spend less), hand off to the financial-forecast skill.

"is my 529 on track" / "how much college will the 529 cover" / "should I superfund the 529" / "roll leftover 529 money into a Roth" / "what do I do with an overfunded 529" → analyze_529_optimization

Always CALL analyze_529_optimization for these — do not answer from general knowledge or quote 529→Roth rollover caps, superfunding multiples, or per-kid college cost figures from memory. When the user gives the numbers (or you have a plan_id), run it and lead with its real output (529→Roth rollover headroom and superfunding moves for the education section), THEN explain. Chain { plan_id } from the assembled plan. Cross-link: pairs with analyze_education_credits (keeping expenses out of the 529 to preserve the credit).

"how much education tax credit can I get" / "AOTC vs Lifetime Learning Credit" / "maximize my college tax credit" / "do I qualify for the American Opportunity Credit" / "should I pay tuition out of pocket or from my 529" → analyze_education_credits

Always CALL analyze_education_credits when the user gives tuition/qualified-expense numbers and income — do not answer from general knowledge or quote AOTC/LLC dollar limits or phase-out ranges from memory; run it and lead with its real output (AOTC vs Lifetime Learning per student, MAGI phase-out, refundable split, and the $4k 529 carve-out), THEN explain. Pass the expense and income numbers — or chain { plan_id } to derive MAGI/filing status. Cross-link: pairs with analyze_529_optimization (the carve-out coordinates with 529 distributions).

"will my estate owe taxes" / "how much estate tax will my kids pay" / "does my state have an inheritance tax" / "am I over the estate exemption" → analyze_estate_exposure

Always CALL analyze_estate_exposure for these — do not answer from general knowledge or quote federal/state exemption amounts or estate-tax rates from memory (they change and are engine-tracked at ~2026 values). When the user gives the numbers (or you have a plan_id), run it and lead with its real output (state-by-state estate / inheritance detail: projected estate, applicable exemption, federal + state tax, effective rate), THEN explain. Chain { plan_id } from the assembled plan's estate section.

"how much life insurance do I need" / "is my coverage enough" / "what happens to my family if I die or can't work" / "do I need disability insurance" → analyze_insurance_needs

Always CALL analyze_insurance_needs for these — do not answer from general knowledge or quote "10x income" or any coverage rule of thumb from memory. When the user gives the numbers (or you have a plan_id), run it and lead with its real output (life + disability coverage breakdown by earner: recommended vs existing coverage, per-earner gaps, monthly disability shortfall), THEN explain. Chain { plan_id } from the assembled plan's protection section.

"what should I do first" / "give me prioritized recommendations" / "what are the biggest wins in my plan" / "turn this plan into action items / next steps" → generate_financial_insights / generate_action_plan

Always CALL generate_financial_insights (prioritized, dollar-quantified insights) and/or generate_action_plan (time-boxed next steps) for these — do not improvise a to-do list from general knowledge; the engine ranks moves by real dollar impact across the whole plan. Run the tool(s) with { plan_id } and lead with their real output, THEN explain. Cross-link: each insight maps back to one of the four plan sections; offer the matching deep-dive route above.

Step 4 — Present the plan

• Lead with headline, then walk the four sections (retirement → education → estate → protection). Keep it one cohesive document, not four disconnected analyses.
• Read back assumptions from assumed_defaults[] (and/or disclosures.key_assumptions) so the user can correct any silent default and you can re-run.
• disclosures.not_advice — relay that this is a planning estimate, not financial advice.
• next_actions[] — each { tool, why, prefilled_args:{ plan_id } }. Follow these server-suggested chains rather than guessing.
• share_url — offer the plan's planfi.app link so the user can open the full interactive plan.

Notes

• All decimals are fractions; all dollars are today's (real) dollars; stock return is real.
• Reuse the plan_id across the session — never re-send the full model.
• The Monte Carlo failure rate may be Calculating…/undefined until the household reaches FIRE at a retirement row — report it honestly, never substitute 0.
• For a FIRE-only deep dive (goal-solving, scenario comparison, "what if" forks, savings sensitivity), use the financial-forecast skill; this skill is the broad multi-section superset.
• Not financial advice. Planning estimates only (approximate ~2026 brackets/limits where tax applies).