7.2: the pipelined relay (fork bb397f99) at 10 bps and a 3-s propagation delay holds 10.01 blocks/s with the
difficulty 7.7 percent over the hash-implied value, the hub taking 7,746 blocks via relay against 4,015 via submit;
the same run at a batch of 1 (the 0.3.17 flow) takes 445 against 801 and reads 0.351. N3 closed in the ledger,
network.md 5.11 and CLAUDE.md's gate line. The 1 bps, 30-s case did not merge on either flow (3 blocks via relay in
40 minutes, no IBD, no error, one reconnect per spoke): at a 30-s delay the blocks in flight exceed the k bound, and
its cause is owed its own reading (7.3).
7.7: the N4 node half's gate (peer-drop.mjs): a still sink dropped its silent peer at 720 s and re-dialled at 750 s;
a moving sink kept the same peer through 155 sinks; the harness's failed shape fails.
7.8 and ledger N6: the stale-block nuisance guard (fork 319da644, f2b25fdf, a22cb261), the shipper's record (310 to
551 connections in an hour from four stale datadirs), the two halves and the knob; the gate (nuisance-peer.mjs, a
node mining stub-engine blocks with proof of work off and restarting on the shared rule with its datadir): the
sender half flat with the skip line, the known-failed shape climbing (3 connections, 2 refusals in 300 s), its
mirror failing, the receiver half flat with one refusal and the ten-minute ban. Also recorded: the vote-or-burn
harness's devnet finality object (--finality devnet) for the replay running to about 05:00Z.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Ledger N5: cargo build -p kaspad alone leaves the igneum-pow feature off (kaspad/Cargo.toml, default kHeavyHash
stub); igneum-miner turns kaspa-pow/igneum-pow on, and feature unification carries it into kaspad only when the two
are built together. A node built alone refused every real block (two fork-gate chains, 0 blocks), the same miner
mined 118 and 115 blocks against nodes built with both packages. Recommendation: igneum-pow a default feature of
kaspad for 0.3.17, the engine named in the version string (igneum_getNodeInfo's powEngine carries it now).
Node document 7.5 and 7.6: vote-or-burn none PASS, the known-failed case FAILS, burn and bonus PASS on the payout
(0.8000 and 0.9000, the pool 0.25 and 0.35) and FAIL on the replay (a node synced from genesis across the span
disqualified 240 blocks as UTXO-invalid: its silence reading differs from a present node's; both switches stay
never until the reading is a pure function of the block's past); the replay gate's own failed shape (another
digest) refuses at the handshake. Fork gate: the rule's known-failed case PASSES (a 32.5 percent key's 171-DAA deep
fork took A's sink 25 s after the heal under the rule as written); the gate-on case found the depth gap (fixed on
the fork, 3301cf32) and then PASSED on the rebuilt pair (B 2,810 bps, 193 DAA deep, heavier by 193, A held, A learned
188 of B's blocks) with a log flood fixed next (9a44fcb8); the harness now waits for each node's RPC and clears
leftover nodes before a case. The chain on the final pair runs; its three lines land in the table when it closes.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
network.md 5.11 and proposal 1: Kaspa's relay requests one round per batch, so at 10 bps and a 2-s RTT the hub took
one spoke block in three (every block a chain block, two thirds of the network's hash merged nowhere); at 400 ms it
took all of them. The fourth gate for any rate step: relay delivers the full block rate at the measured inter-region
RTT with margin. CLAUDE.md's block-rate line carries it. The live 1 bps devnet offers under one block per RTT and is
nowhere near it. Controller rows (node document 7.2): rule v3 at 600 ms holds 9.93 blocks/s with the difficulty 4.5
percent over the hash-implied value; the live rule v2 reads 22 percent under on the same DAG shape.
Ledger N4: the vote weigh-up's silent desktop key, upload by upload from miner_logs and the observer: the voter never
died and --no-vote was never set; the node fell off the network (a digest-mismatch refusal, then a same-digest peer
that relayed nothing), its tip froze for 53 minutes, and the miner hashed and voted on the frozen template with
synced=true throughout while the node accepted 1,750 blocks into a dead branch. The fix pair on the fork (miner
e6e1fbe2, node 96619b7d); infra/fast-time/miner-stall.mjs: stall PASS (STALLED at 30 s, exit 45 at 31.5 s), moving
PASS, today's miner FAILS as it must.
infra/fast-time/fork-gate.mjs and vote-or-burn.mjs with the replay gate (--replay). First lines: vote-or-burn none
PASS, the known-failed case FAILS, burn PASS on the payout (ratio 0.8000, pool 0.25) but the REPLAY FAILS: a node
synced from genesis across the span disqualified 240 blocks as UTXO-invalid (its silence reading differs from a
node that was present), so both vote switches stay never until the reading is a pure function of the block's past;
fork gate: the first run idled the honest side and the depth read 0 (fixed on the fork, 3301cf32: depth along both
chains), the re-run's lines follow. The tree ships as 0.3.17 (node document section 7).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The ledger's header and submission lines, G3's status line, and the fixes file's legend, row 9, decision (b) note and section 5 record the four decisions. Two literal pattern mentions reworded so the identity check passes with both files on its export list.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Report docs/review/redteam-2026-10-04.md: finality attacks s1-s8 plus 34% withholding, 50/50 long partition, F23 and
F24 custom runs, ordering harness, execution suite and EVM smoke, proving hostile tests and a proof flood, difficulty v2
timestamp forging in the simulator. New fails: the fast-time harnesses corrupt the u64::MAX sentinels of the override
(F25), a block or transaction flood grows the node by hundreds of MB in a minute (M30), the coinbase does not fit the
204-byte limit on mainnet, testnet and simnet parameters (M31). F23 and F24 reproduced on this build; F21's bound
measured at one window of the side's own DAA. Scenario scripts under tools/finality-attacks/redteam/.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
attacks.py scenario 2 under rule v2 and Kaspa's DAA: a pulse never earns more blocks per hash than steady mining
(weight per hash 0.26 and 0.98); the economy simulator at the devnet's 124 MH/s; finality_sim scenario A (the
window is full on day 35 to 41 from zero history). The first fast-time s5 attempt failed on an override field the
integration build does not know; the re-run on the finality-fixes build is queued.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Spec 06 rows O-3.1, O-3.14, O-5.9 and O-5.11 carry the sweep's evidence (fix row 124 done). M20: the devnet's pruning
point leaves genesis between DAA 108,000 and 151,200, about 14:00 UTC 5 October to 01:00 UTC 6 October, after which a
fresh node rejects the honest pruning proof under the stub. The ledger ends with the sweep's status-update section;
the review file carries draft counts and the three findings.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Status lines updated from the bench-log, sim/results_v2.md, sim/difficulty/attacks and sim/economy where the
evidence existed and the ledger still said Open (M15, M17, M19, M24 fixed and live; F1, F7, F19, E12, E15 answered
with evidence; M26, M27, X21 fix built on miner-reliability; the rest annotated with what the experiment needs).
New: sim/economy/security_budget.py (E15 fee grid, price paths, hashrate response), fud-fixes section 2.5 (rows
117 to 126), docs/review/ledger-sweep-2026-10-05.md (the running table).
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Conflicts resolved keeping both: engine.rs carries the sweep state (miner-eff) and the node watchdog (reliability);
the STATUS line goes through watchdog::parse_status and still feeds the sweep's rate sample; main.rs declares both
modules; bench-log.md keeps both entries. site/build.mjs keeps master's partial-injected pages and adds the /miners
bench table through the same page() with active: 'miners'; the Miners link is in site/partials/nav.html; sitemap
gains /miners; every generated page rebuilt with node site/build.mjs.
docs/bench-log.md: the fake-worker measurements (slow start one trip and 2.0 s restart, fake-fast guard in under
0.1 s, exit 43 at 8.8 s, CPU re-check stop at 0.5 s, stall guard at 60.1 s with STATUS lines through the silence,
one prepare per epoch with refused retries held; app: zero-rate restart at 79.6 s and faulted at 75.4 s on the
repeat, no-status restart at 90.4 s, silent node restarted at 150.7 s and synced 7.2 s later; no double restart on
the miner's own worker restart). Two defects the harness found are named with their fork commits.
docs/fud-ledger.md and the round-4 review table: M26, M27, X21 Fixed with the commits.
engine.rs: the miner's restart note no longer hides the fault reason on the card.
tools/reliability: the harness matches the miner's stderr lines where they are printed there.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Conflicts resolved keeping both: Settings and SettingsState carry the sweep fields (miner-eff) and dev_fee /
fee_total (dev-fee); the engine's settings snapshot sets both and the DevFeeState line stays.
Decision of 4 October 2026 (evening): the Igneum Miner software takes a visible, switchable 1% dev fee, the norm
for GPU miners; the protocol stays fee-free. The miner side (--dev-fee, the 1-in-100 template counter, the audit
command) is on branch dev-fee of the node fork.
- app: settings.dev_fee (default on) passes --dev-fee 0 to igneum-miner when off; Settings shows the miner's own
"dev fee 1% (1 block in 100) to 0x..." line next to the rewards address with a switch; the engine parses the
miner's start line and its dev-fee block lines (fee_session, lifetime fee_total, an event per fee block)
- packaging/hive: h-manifest.conf, h-config.sh, h-run.sh, h-stats.sh, make-hive-package.sh (igneum-hive-<v>.tar.gz
with the Linux igneumd, igneum-miner and both GPU workers), README with the Flight Sheet, selftest.sh (bash -n,
stub binaries, the three hooks the way Hive runs them, the stats JSON parsed). Hive itself is untested
- infra/cross/build-workers-linux.sh: the NVRTC and OpenCL workers cross-compiled for Linux with zig;
proto-opencl/cl_dynamic.h gains the Linux dlopen branch (libOpenCL.so.1)
- tools/dev-fee/run.mjs: the fee-block test network (two nodes on 29900+, three CPU miners, payouts audit)
- docs/design/miner-dev-fee.md (mechanism, flag, lines, the DEV_FEE_ADDRESS placeholder and the devnet address),
docs/fud-ledger.md E18 and the E5/L9 status line, litepaper "What a miner's hour looks like" paragraph, homepage
miner section note
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The intake key sits in every miner package, so the relay now lets it report only (drop text and files, ack, done,
register, upload). Posting a run or task, or renaming and re-roling a machine, needs the console token.
The prove host wrote proofs through SP1's unbuffered save: on WSL2 under /mnt/c the 18 MB core proof of a shard
took longer to save than to prove. Proofs now go through a 4 MB buffer with a timed 'saved' line, and
prove-shard.sh keeps results on the Linux side and copies them per stage. Ledger P20 updated.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Live devnet v4: a second RTX 5090 joining 7 minutes into an epoch left the whole-epoch reference lane polluted for the hour; the short lane read 11 to 25% above it and the 25% trigger flipped between the two for 40 minutes (102M to 164M, 54 to 81 blocks a minute). Record and hash-rate truth under sim/difficulty/records/. sim.py gains a DAG model (miners on nodes with igneum-miner's template staleness, GHOSTDAG, the rule as the node runs it) and --live replay: std of log difficulty 0.115 against the record's 0.134, 4.3 peaks of 1.31x against 4 of 1.37x. The brief's candidates (short lane 240/360, ease clamp 3%, clamp once per DAA second, hysteresis, median of three) leave 0.09 to 0.13; capping the reference lane at the newest 600 blocks of the epoch gives 0.026 with no flips. Rule v2 = that cap, epoch lane only, behind difficulty_v2_activation_daa (devnet-v4 fork). Attack suite and synthetic set before and after, 3-node test network of the switch (testnet_v2.py), analysis document, spec 2.3, bench-log entry, ledger M24, fast-time file carries the new field.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
A lock needs two thirds of all 30-day weight signing; finality pauses
whenever less than two thirds is connected and signing, the chain runs
on proof of work meanwhile and the node reports it. Spec 3.3, 3.3.1,
3.7, 3.9, 3.10 Q3 row, 3.11 rewritten with the new arithmetic (safety
one third in every view, liveness two thirds connected, the per-view
window bound stated as 3.7 item 9); O-3.15 decided, O-3.16 closed,
O-3.18 and O-3.19 narrowed. Simulator: --floor, the +local partition
mode, scenario L; A to L re-run at the 2/3 floor over five seeds with
the 0.85 deltas in results_v2.md. Litepaper finality sentences and the
'does not claim' item. Ledger F2, F9, F16, F18 restated, F21 added
(the window bound and the post-heal finality fork from the devnet).
Bench-log: node build and tests, simulator deltas, three-node six-voter
runs of 6A, 6B and 6A with a long heal on ports 29200 and up.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
igneum-pow 0.2.0: generator v2 draws exactly 16 load slots from instructions 1..63, a
load's source from the registers written earlier and not read by a load since, the other
48 ops from the ten non-load weights; accept.rs is spec 01 section 1.4.6 (static: no
stale load source, every register injected; dynamic: 64 units on the seed-keyed
closed-form dataset, no constant bit, no lane-constant site, under 164 saturated, bias
within 136 of 1024, distinct addresses above 245,760); a rejected candidate is replaced
by the next attempt of the seed (seed || k_le32), 32 a consensus fault. Packs carry the
generator version, attempt and program id. Version 1 kept as generate_v1 for the census.
Packs: igneum-genesis, igneum-hourly, igneum-genesis-mh regenerated by igneum-pow export;
new igneum-devnet-v4-epoch0 (devnet genesis hash, day bytes 20730). Checks: Rust 39 of
39 tests; Metal natively via the Swift port (export cross-check 3 of 3 warps, identical
programs and vectors on five seeds incl. three with attempt 1, fuzz 2,000 of 2,000);
CUDA emu 4 of 4 packs; OpenCL emu 2 packs x 2 configurations; Apple OpenCL 4 of 4 packs
at 27.9 Mhash/s. Census 20,000: 5.225 percent rejected, accepted distinct mean 127.887.
Spec 01 0.2 (1.4.2, 1.4.3, 1.4.6, 1.11, 1.15, 1.16, 1.17), igneum-pow README, the CUDA,
OpenCL and Metal test notes, bench-log entry, ledger M5 and M6 Fixed.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
