feat(contracts/ci/docs): property testing, fuzz targets, runbooks, WASM size tracking

[Fidelis900]

Summary

This PR resolves four issues across contract testing, operational docs, and CI
transparency in a single branch.

---

#999 — Property-based invariant testing + INVARIANTS.md

• Added proptest as a dev-dependency to contracts/predict-iq/Cargo.toml
• Created src/modules/property_invariants_test.rs with 5 proptest properties:
  • Prop 1 — Stake conservation holds after arbitrary bet sequences
  • Prop 2 — Cancelled status is terminal; new bets are rejected
  • Prop 3 — Resolved status is terminal; new bets are rejected
  • Prop 4 — total_staked is never negative
  • Prop 5 — Full cancel + refund drains total_staked to zero
• Created contracts/predict-iq/INVARIANTS.md documenting all 7 financial and
  state-machine invariants with Certora Prover and KEVM exploration guidance

---

#1000 — cargo-fuzz targets for top 3 entry points

• Created contracts/predict-iq/fuzz/ with a proper cargo-fuzz Cargo.toml
• Three libFuzzer targets:
  • fuzz_place_bet — fuzz outcome, amount, and timestamp inputs
  • fuzz_resolve_market — fuzz market ID and winning outcome (including out-of-range)
  • fuzz_withdraw — fuzz market ID on a cancelled market
• Added contract-fuzz CI job to .github/workflows/test.yml:
  • Runs each target for 60 seconds with -max_total_time=60
  • Uploads crash artifacts on failure
• Documented fuzzing setup, targets, corpus/crash handling, and CI in
  contracts/predict-iq/README.md#Fuzzing

---

#1001 — Operational runbooks for incident response

Five new runbooks under docs/runbooks/:

File	Scenario
api-outage.md	API completely unreachable — ECS checks, ALB triage, force redeploy
redis-failure.md	ElastiCache down — failover, memory/connection diagnostics
email-queue-backup.md	SQS queue depth spike — worker restart, DLQ inspection and replay
stellar-rpc-unavailable.md	RPC unreachable — fallback switch, ledger lag measurement
ecs-task-crash-loop.md	Task crash-loop — exit code guide, rollback playbook

Each runbook covers: detection signals, immediate mitigation, investigation
steps, escalation path, and post-incident actions.

Five new Prometheus alert rules added to performance/config/alerts.yaml
(APIOutage, RedisFailure, EmailQueueBackup, StellarRPCUnavailable,
ECSTaskCrashLoop) each with a runbook_url annotation.

---

#998 — WASM binary size budget transparency

• Added "Print unoptimized WASM size" step to the build-optimized CI job
  so the raw pre-optimization size is logged on every build alongside the
  enforced optimized-size check (existing WASM_SIZE_LIMIT_BYTES = 65536)
• Added WASM Size Tracking table to CHANGELOG.md [Unreleased] section
  with budget rationale and measurement instructions for contributors

Note: The 64 KB budget and WASM_SIZE_LIMIT_BYTES reference were already
documented in contracts/predict-iq/README.md prior to this PR.

---

Closes

Closes #999
Closes #1000
Closes #1001
Closes #998

[drips-wave]

@Fidelis900 Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits