feat(quest-engine): deliver uncapped staking multiplier payouts

[Birdmannn]

fix(quest-engine): deliver real boosted payouts on Build Quest approvals

Problem

The staking multiplier was being fetched in review_submission but the final
learner payout was silently capped back to base_learner_amount. A learner with
a 1.2x multiplier received exactly the same amount as an unstaked learner,
making staking incentives ineffective for B2B bounties.

Solution

Redesigned the approval payout in review_submission with an explicit two-source
funding model:

Source	Covers
Quest escrow	fee (15%) → RewardPool + base_learner_amount → learner
RewardPool	boost_delta (boosted_amount − base) → learner

The quest escrow budget check (remaining >= quest.reward_amount) is preserved
unchanged. The boost delta is drawn from the RewardPool via distribute_reward,
which fails deterministically if the pool lacks sufficient balance — making the
funding constraint explicit and observable.

Basis-point math is unchanged: boosted = (base * multiplier) / 100.

Changes

contracts/quest-engine/src/lib.rs

• Removed the cap that forced learner_amount back to base_learner_amount
• After transferring fee + base from escrow, conditionally calls
  reward_pool.distribute_reward(contract, learner, boost_delta) when
  boost_delta > 0
• consumed_amount continues to track only what is drawn from the quest escrow
  (quest.reward_amount), keeping budget accounting consistent

contracts/quest-engine/src/test.rs

• Added MockStakeVault200 — a 200x (2.0×) multiplier vault for tier testing
• Added MockRewardPoolTransfer — a mock RewardPool that actually transfers
  tokens, needed to verify end-to-end learner balances for boosted payouts
• Added setup_with_boosted_multiplier() — wires MockStakeVaultWithMultiplier
  (120x) + MockRewardPoolTransfer into a full QuestEngine instance
• Rewrote test_review_submission_with_120_multiplier to pre-fund the pool for
  the delta and assert the learner receives the real boosted amount (1020, not 850)
• Renamed test_refund_after_capped_multiplier_returns_leftover_only →
  test_refund_after_boosted_multiplier_escrow_is_fully_consumed with updated
  assertions matching the new accounting model

New acceptance-criteria tests:

Test	Covers
test_staked_learner_receives_more_than_non_staked	120x learner earns more than 100x for the same quest
test_basis_point_math_for_100_120_200_multipliers	Correct integer math for all three multiplier tiers
test_review_submission_with_200_multiplier_draws_delta_from_pool	200x tier end-to-end payout
test_review_submission_fails_deterministically_when_pool_cannot_cover_boost	Contract panics when pool has no balance for the delta

Payout walkthrough (reward = 1000, multiplier = 120)

fee            = 1000 * 15 / 100        = 150   → RewardPool (from escrow)
base           = 1000 - 150             = 850   → learner    (from escrow)
boosted        = 850  * 120 / 100       = 1020
boost_delta    = 1020 - 850             = 170   → learner    (from RewardPool)
─────────────────────────────────────────────
learner total                           = 1020
escrow consumed                         = 1000  (fee + base)

Test results

running 59 tests
... all pass ...
test result: ok. 59 passed; 0 failed

Full workspace: 202 tests, 0 failures.
Closes #87