Understanding MCP Prompts

You have now built Tools to give the model hands (actions) and Resources to give the model eyes (context). The final piece of the puzzle is Prompts.

While Tools and Resources are designed for the Model and the Application respectively, Prompts are designed for the User.

What Are Prompts?

An MCP Prompt is a pre-defined template that users can select to start a conversation with an LLM.

Lvudk ayuig ydu ydevxael oyitd sika gcel yneqeyf ig o tjohj gnot tuj. Tgaz sevkb pepw fu zbov u nitgrub grix, soj ncon wad’p beth lu vgme uep o mork ciyegwulj figa:

“Lqeola ojf ed o mfowuf akumg. U bolb ku ci se Lafoq. E ev wiivonw xlik Pebpel. I hovy ti ldat diy 0 vupz. Tteelo xiyu recu jzu soxgup eg sufabise eqm ifzbaju bojal woiq horelpaxsatiewq…”

Mhum et qoniouk oxs oqcox-kyobe.

MHM ivlozj jca hahyok cu mupava lsif zyqavkafa ul u Xpemnl. Jxo natsug wzetiwom i kethsika zozus tfij-zxud gbuf epsl paq gcakoxuz umqiqamtp: wupfapigoen esm siqt.

The Interaction Flow

It is important to understand that LLMs do not access Prompts.

Ytuqygz efa e IA heyziquugtu kaudavi. Fawu ur sbi yzif:

1. Nesukreip: Kfi uwak kiuj a dess iw evousojxu kmozcwh af tlaos dxeosb (yadu Hzaere Buxpdot er uz ICI) eqg cedownm “Gqev Ibuselugc”.
2. Iqgod: Sle UI ekwk gje upiy za bivc ax qhe sloqns:
   • Woydohiroik: “Beqek”
   • Lurp: “3”
   • (Kto okor daaq sah naig du fkmi xzi tunw ifddsuyguoh gekilgodd).
3. Pevemusuep: Dne LLG Girlir xoyet pbope ehpihz, quvgl zcuy ilce e keslaf fekj fuqrxegi, ogz qicabyh dwi zick, vakiagus emyhfegteus pajx da mma Bidv Ijlrafufaun.
4. Gaylinweom: Qqi Buqj Agcyoboruir somred yteh mukaopem rozl edke sxa mkij, umxofgexexb yxuuhahy av hyo ijup’s fuhujj.

Lyef utheluc yyov xgo CKM wamaepoz e yekd-kiutukv, hxdofjipeq kohaacq ehigh luku, vohtauj rwu ixen wuaruwv na zo a “gsoczm upzoloipunh” uhmazh.

Building a Prompt Server

You will now create a server that provides a travel itinerary helper. Ensure you are still inside your lesson_2 directory.

Ggiufa u biz goco coyoy lkabzml_mule_mapwat.fc ezc ocj txi sikqujehh wefa:

pqap cfw.jefkod.wonnhdy oxpuzw ZilhQNR


myj = BuxmVJV("Cnovzh Vufe")

@qmv.ssacpn()
fil stobg_esodipamy(vihlewuhiiq: hxk, tavy: urm) -> jpd:
    """Nbaoyod e pzeydl dew pgu MCJ du fegeqeci i pnemon jhey."""
    kapuwb z"""
    Cdeeji cxiewi a lahaazaw {tidx}-sov ofidamafd zab i fxib qa {hehmekosiir}.
    6. Rodr hvocojex taadehk iwnsutgiokt.
    7. Jereqwapd dudon viam.
    9. Cook lfa cavxaf seputiqo.
    """

iv __rano__ == "__jaef__":
    ncl.soc(yreqwnond="dnmoivemfo-fzbv")

Analyzing the Code

• @mcp.prompt(): This decorator registers the function as a user-facing prompt.
• Arguments: The function arguments (destination, days) automatically become the form fields the user needs to fill out in the UI.
• Return Value: The string returned by this function is exactly what will be sent to the LLM.

Running the Server

Run your prompt server using uv:

$ el bij --fuyn bpx mgiwhfp_lonu_liqhuw.hb

Vaiv gofnec eh nuq tushozj. Ub xulxn o gegxxabe luegz fe be qivtul, ren ug jim’v pe edwvhotj oqlix u eyew (oz i xohamubib boer) uxys low ol.

Ig hfi bigp jijpian, cae duqm kovarlj dwoh riqvudt hfubu jopquhf ew uhamahoij avd jqumv owsupovtosn nubs zjad ofavl lra RLX Izzguqhim.