Notes / Audits and snapshots
Source: chains/tempo/data/scored_targets_2026-09-17.md in the repository, rendered here without changes. Source file
Promotes 0 section 4.1's rules (R1-R7) into the standard scorers.py / score_all() shape every other ecosystem in this project uses: 3 . The underlying read-and-score logic was already fully implemented and live-tested in scripts/methodology_test.py on 2026-09-16 -- this pass adds only the score_all() sequencing that file's own main() did inline and unexported.
| Target | Address | adminKey | multisig | timelock | oracleAuthority | crossExposure | composite | chainCappedComposite |
|---|---|---|---|---|---|---|---|---|
| Tempo L1 validator registry (ValidatorConfigV2, chain baseline) | 0xCCCCCCCC00000000000000000000000000000001 | 50 | 37 | 0 | 100 | 100 | 31 | n/a (this IS the baseline) |
| USDC.e (Bridged USDC, Stargate) | 0x20C000000000000000000000b9537d11c60E8b50 | 65 | 98 | 0 | 100 | 100 | 55 | 31 |
| USDT0 | 0x20c00000000000000000000014f22ca97301eb73 | 65 | 58 | 0 | 100 | 100 | 43 | 31 |
| pathUSD (native, added 2026-09-17) | 0x20c0000000000000000000000000000000000000 | 20 | 0 | 0 | 100 | 100 | 8 | 8 |
The first three match data/methodology_test_2026-09-16.md's hand-derived numbers exactly. pathUSD is new the same day, added as a follow-up -- see below.
pathUSD is Tempo's own native/first-party stablecoin (the fee-fallback token cited in METHODOLOGY.md section 5), NOT bridged via LayerZero (not listed among section 3.7's bridged assets) -- score_token() was extended to accept lz_oapp=None for exactly this case (see its own comment in scripts/methodology_test.py) rather than assuming every TIP-20 is bridged. Second-largest stablecoin on Tempo by supply (~$39.3M live, bigger than USDT0's ~$9.7M), promoted for the same reason io was added to Hyperliquid the same day: real usage, not just being next on a list.
Root-control set, live-read via a full-history RoleMembershipUpdated scan: DEFAULT_ADMIN_ROLE is a single EIP-7702-delegated EOA (0x79C6631F...4a4E, k=1,n=1); ISSUER_ROLE is a 100-byte contract (0x8354D80E...9058) that does not resolve via any of owner()/admin()/ minter()/controller()/vault() -- an "unresolved contract" per rule R2, the weakest possible key by rule R3's own definition. This -- not a degraded/failed read -- is what drives adminKeyScore=20, multisigScore=0: the token's own genuinely weakest authority path could not be resolved to a named signer set at all, a materially different (and weaker) shape than USDC.e's 5-of-7 OneSig or USDT0's 3-of-5 Safe. compositeScore=8, the lowest of any Tempo target scored so far, including the chain baseline itself (chainCappedComposite=8, no effect since pathUSD's own score is already the binding constraint).
Every eth_call fact this scorer reads is cross-checked against a second, independent RPC (https://tempo-rpc.publicnode.com alongside the official https://rpc.tempo.xyz) INSIDE methodology_test.py's own call2() helper -- this isn't an optional pass run separately, the read primitive itself raises RuntimeError on any disagreement between the two endpoints, so a passing run is a cross-RPC-verified run by construction. This is a stronger guarantee than the same-day Solana promotion could get for one of its two targets (rate-limited on retry) or Hyperliquid (no independent second source exists for HyperCore at all).
Role-holder enumeration for both tokens comes from a full-history RoleMembershipUpdated log scan (100,000-block windows, the official endpoint's limit) -- https://tempo-rpc.publicnode.com refuses this bulk eth_getLogs query (HTTP 403, a documented, disclosed single-source gap already flagged in the methodology test itself), so the SET of candidate holders comes from one source, though each candidate's actual role membership is then confirmed with hasRole on both endpoints. A role grant missing entirely from the official endpoint's logs would not be found by either endpoint -- carried forward as an open item, not silently treated as fully double-sourced.
USDC.e is the largest stablecoin on Tempo by supply (~61.6M) and exercises every TIP-20-specific part of the methodology (role system, TIP-403 policy, a bridge contract as issuer via a LayerZero OneSig multisig). USDT0 tests a different issuer design (Tether's own proxy-based OFT, not Stargate, with a live custom blacklist policy). pathUSD (added same day) is the weakest authority shape found on Tempo so far -- see its own section above. The L1 validator registry is scored because it is the authority ceiling for every other Tempo target (chainCappedComposite, METHODOLOGY.md rule R6b) -- the same role the Hyperliquid L1 and Robinhood Chain's rollup-authority target play in their own ecosystems.
getActiveValidators() returns registry entries (14 active on mainnet at last read), not the current DKG committee. Resolved, not just open, as of 2026-09-17: TIP-1070 (the spec that would define a getCommitteeMembers() read) is status: Draft, protocolVersion: TBD, and absent from the live mainnet fork schedule (tempo_forkSchedule, active hardfork T11) and from the next hardfork's own meta-TIP -- the committee-state precompile does not exist on Tempo mainnet yet at all, so registry size is not a stopgap for an undiscovered address, it is the only thing currently readable (METHODOLOGY.md section 7, open point 3).data/methodology_test_2026-09-16.md but not folded into any tracked target's crossExposureScore, which only counts overlap among Tempo targets scored in the same run (METHODOLOGY.md rule R6, matching this project's existing within-ecosystem convention elsewhere).python3 -c "
import sys; sys.path.insert(0, 'chains/tempo')
from scorers import score_all
for r in score_all():
print(r['target'], r['compositeScore'], r.get('notes'))
"
Runs about 2 minutes, mostly the full-history log scan (about 400 100,000-block windows against the official endpoint).
Not an audit, not investment advice. All notes ยท Back to the dashboard.