Realised pronouns are a bit of a pain, so I'll leave them in the workshop for a bit. Here are the numerals, almost as painful due to their poor attestation.
1. The number one is unstressed but the likely form is úíns, and might have followed the pronominal declension as Latin ūnus did. The ordinal number would be either prismús (would the suncope of ú be blocked by the cluster or what? Maybe become prisem, prismú, prismúm) or prúms depending on whether it follows Paelignian or Umbrian, more likely the former as it's more closely related to Oscan.
2. The number two is also unattested. The form may be duus, duvas, following the plural of o-/ā-stems rather than continuing the dual as in Latin. The word altter 'other' could likely be used as an ordinal, but it may have had the form dutis, dutiú, dutim as well. Where Latin has bi-, Osan most likely had di-.
3. Oscan trís (although it should be tríís) is attested as the masculine and feminine accusative. It has the root treí- and follows the plural i-stem conjugation. The ordinal would be tertis, tertiú, tertim.
4. We run into difficulty with four as Wallace and Buck disagree. There are the forms pettiur (Wallace), petora (Buck) and petirú- (both). Wallace's forms could be in the same relation to each other as the n-stem nouns, with -i- in place of -iu- in oblique forms. Buck's form is however easier to explain from the PIt form *kwetwor. Wallace would require a form *kwetur, with samparasana of the *w after syncope, u > ju after a dental and then lengthening of a voiceless stop before /j/. All in all possible but not as likely. Buck also gives a possible cardinal trutum which believe it or not is explainable as a zero form of the PIE numeral, *kturīya- > *ktru-to- (in which case it would be turz, turtú, turtúm.
5. From here on the cardinals are indeclinable (not sure about ordinals). Five is púmpe and its ordinal is púnz, púntú, púntúm.
6. The cardinal might be seks (or should it be sííks?) and the ordinal sest- (is /sts/ a valid final cluster? I need a way to resolve awkward consonant clusters)
7. There is no evidence for the number seven, but it might be of the form seftúm, with the ordinal sefz, seftú, seftúm.
8. The cardinal is probably úhtu and the ordinal úhtavis as in the name, although I don't know how likely that is.
9. The ordinal is probably núvems, núv(e)mú núv(e)múm or something similar, following the cardinal rather than the Latin nōnus?
10. The cardinal is probably dekem, with the ordinal dekems, dekmú, dekmúm although there is also a possible form deket-.Edit:
Here's a little chart
úíns, úínú, úínúm ... priísmús, priísmú, priísmúm
duus, duvas, duvú ... dutis, duttiú, dutim
tríís, triiú ... tertis, tertiú, tertim
pettiur, petirú ... trutis, truttiú, trutim
púmpe ... púnz, púntú, púntúm
sess ... sestús, sestú, sestúm
seftúm ... seftús, seftú, seftúm
úhtu ... úhtavis, úhtaviú, úhtavim
núvem ... núvems, núvmú núvmúm
dekem ... dekems, dekmú, dekmúm
Unresolved questions: would *prīsmos, *sekstos, *septos keep the o because of the cluster or deal with it in some other way? Also, what about the word dekez, deketú, deketúm?
Would duttiú and truttiú lose or generalise the double tt or just leave it as is? The doubling of consonants before /j/ (and sometimes before /r/ and /w/) seems to be a later development and possibly still productive by 100BCE.