docs(v5-next): backport v5 testnet connection details + evergreen doc fixes

[AztecBot]

v5-next docs backport covering the remaining survey candidates, the partial-notes docs from #23564, and the slashing-configuration rewrite from #24005 (now adapted to v5). Complements #24242 and #24243.

Files

v5 testnet connection details

• getting_started_on_testnet.md — NODE_URL → https://v5.testnet.rpc.aztec-labs.com, SPONSORED_FPC_ADDRESS → 0x261366…7880, dropped the stale "~36 seconds" block-time row.
• docs/docs/networks.md — testnet column → v5.0.0-rc.1 (rollup 4239416255, RPC, L1/L2 addresses). The values getNodeInfo returns were verified against the live testnet node; the slashing stack (Slasher, Tally Slashing Proposer) was verified on-chain via rollup → getSlasher() → PROPOSER(). Mainnet column: 4.3.0 → 4.3.1 (rollup version 2934756905 unchanged, confirmed on the live mainnet node). Also corrected the governance table's Slashing Round Size to 4 epochs (128 slots) on both networks (was 128 / 64 epochs, which matched nothing in config); confirmed on-chain on both rollups (ROUND_SIZE_IN_EPOCHS = 4).

slashing-configuration.md (from #24005, adapted to the v5 testnet)

Brings in #24005's structural rewrite (correct ejection mechanism, 72s slot math, grace period) and corrects the network-specific numbers to the v5 testnet config:

• slot 72s; round size 128 slots (4 epochs × 32) ≈ 2.6h
• slash amounts 100,000 / 250,000 / 250,000 (AZIP-16 full-stake preset) vs the old page's flat 2,000
• local ejection threshold 199,000 (so any single offense ejects a validator that joined at the 200,000 Activation Threshold) vs the old "98% / 3 slashes"
• execution delay 2 rounds (~5h) on testnet vs the old "28 rounds / ~3 days"
• grace period stated for both networks: testnet 64 slots (~1.3h), mainnet 1,200 slots (~1 day)

The old v5-next page was materially wrong for v5 (36s slots, 1%/3-strikes model). All figures were confirmed on-chain against the live Sepolia rollup (0xe439… → slasher 0xd9f4… → proposer 0xaB9B…): slash 100k / 250k / 250k, local ejection 199k, activation 200k, protocol ejection 100k, round size 128 slots / 4 epochs. The mainnet contrast figures (190k local ejection, 2k/5k slash, 28 rounds) match spartan/environments/network-defaults.yml.

Post-review fixes in this branch: restored the v5 aztecAdmin_getConfig method name (nodeAdmin_* is the legacy pre-v5 namespace, per migration_notes.md), and replaced the single mainnet-only grace-period value with both networks' values.

Partial notes (from #23564, still open on next)

• partial_notes.md (expanded), partial_notes_as_payment_endpoints.md (new), docs-words.txt (unlinkable). All cross-links and the #include_code/references paths verified present on v5-next. The note-hash ordering (H(storage_slot, partial_commitment, value)) and validity commitment (H(partial_commitment, completer)) match uint_note.nr / note/utils.nr; the AIP-20 functions (initialize_transfer_commitment, transfer_private_to_commitment, transfer_public_to_commitment) are documented on v5-next's aip-20.md and defined in the standard's reference implementation.

Evergreen

• community_calls.md — refreshed meeting times / links / calendar.
• ai_tooling.md — "never call bb directly" agent guidance (doc prose only).

Sources: v5 release-docs flow (7d6b320, f5fadef, 8bfd77d, b84af54, 80830d0), 8aed7c5/22393fe, 9ebff63, #24005 (slashing), #23564 (partial notes). No tracked issue is closed by this backport.

[AztecBot]

Flakey Tests

🤖 says: This CI run detected 1 tests that failed, but were tolerated due to a .test_patterns.yml entry.

\033FLAKED\033 (8;;http://ci.aztec-labs.com/5039f193fa054832�5039f193fa0548328;;�):  yarn-project/end-to-end/scripts/run_test.sh simple src/e2e_epochs/epochs_optimistic_proving.parallel.test.ts "handles a reorg arriving while the top of the epoch is proving" (96s) (code: 0) group:e2e-p2p-epoch-flakes