fix(scripts): sync cost model genesis values with mainnet

[OlufemiAdeOlusile]

Summary

• Genesis specs (genesis.conway.spec.json, all 3 variants) still carried the Conway-launch PlutusV3 length (251 entries); local_slow's cost_models_list.json was already ahead at 297. Mainnet has since raised cost models via governance (V1=332, V2=332, V3=350 as of epoch 640).
• common.sh and common-start-fast/common-start-slow additionally truncated PlutusV2/V3 cost models back down to the old lengths on every cluster start via jq, silently discarding correct genesis values. Removed those truncations from common code entirely (not re-added anywhere).
• Removed a redundant extraConfig re-injection jq block from common-start-fast per review feedback - the extraConfig field itself is untouched and still present in all three variants' genesis spec files (it was always inert; only the jq that copied it into the generated genesis output is gone).

Per-variant end state (verified by actually running each)

• local_fast / mainnet_fast: genesis only, no extra jq/code - {PlutusV1: 332, PlutusV2: 185, PlutusV3: 350}. PlutusV1/PlutusV3 pass through the spec unchanged; cardano-cli genesis create-staked hard-caps PlutusV2 at exactly 185 regardless of spec/extraConfig content, confirmed by isolated testing against two different cardano-cli versions (10.11.1.0, 11.1.0.0).
• local_slow: genesis intentionally stays at the original historical values - genesis.alonzo.spec.json is byte-identical to master, not touched by this PR net of all commits. PlutusV1 stays at 166 throughout (genesis and cost_models_list.json both 166, never changes). PlutusV2 starts at 175 in genesis and grows to 185 through the historical era-transition proposals (Alonzo-PV6/Babbage), then the existing PV9 governance flow (_update_cost_model_in_pv9 / _wait_pv9_cost_model_enacted, using cost_models_list.json) explicitly asserts it lands at 185. cost_models_list.json's PlutusV1/V2 arrays are the original historical bytes (master's own values); only PlutusV3 was updated (297 -> 350), matching genesis.conway.spec.json byte for byte.

Open question, not resolved in this PR: reaching PlutusV2=332 (full current-mainnet value) on local_fast/mainnet_fast would require a governance action submitted after genesis (the same pattern local_slow's PV9 test already uses) - extraConfig does not do this; nothing reads it. A governance-action mechanism was added and tested (confirmed working), then reverted after re-reading the review feedback as asking for jq removal only, not new logic. Whether to add it back is a decision for review, not made in this PR.

Test plan

• make lint passes
• local_fast: cardano-cli query protocol-parameters reports {PlutusV1: 332, PlutusV2: 185, PlutusV3: 350}
• mainnet_fast: same result, confirmed
• local_slow: full Byron -> Conway startup completes, PV9 cost model proposal enacts, hard-fork to Conway PV10 succeeds; final state {PlutusV1: 166, PlutusV2: 185, PlutusV3: 350}
• Full cardano-node-tests pytest suite re-run against the current (post-revert) {332,185,350} state: test_pparam_update, test_pparam_keys, 8xtest_legacy_proposal_submit, test_incompatible_cost_models, test_node_upgrade.py::test_update_cost_models - all pass