Claims campaigns executed — 2026-09-07
Owner: CODEX-INTEGRATION. Evidence level: real ELF ProgramTest, in-process bank. No local-validator or devnet execution is claimed by these results.
Source and execution
Live repository: /Users/ember/dev/dclutch. Measured checkout: hbox:/tank/dregg-build/dclutch-codex-20260907. Accepted source: cac1a0593cba45ad1b29e565e933aac865e1f717, detached and clean for the build and run. Both repository root and HEAD were printed with the run. No manifest or lock change was needed; locked workspace metadata succeeded.
The full suite runner performed the locked package check before the SBF builds, verified the Token-2022 fixture against its provenance, and freshly built its nine first-party program/caller links. Execution was contained by swarm-build on hbox, with SWARM_MEM_MAX=32G, CARGO_BUILD_JOBS=6 and serial test threads. Neither stack-diagnostic shape nor the scheduler's silent no-op was admitted.
Run root: /tank/dregg-build/dclutch-codex-evidence-20260907/committed-runner/run.e0CwMY. The runner was invoked without target selection, which selects these six explicit targets; it never invokes an unfiltered package suite.
| Test target | Passed | Failed | Runner exit |
|---|---|---|---|
| claims_founding | 7 | 0 | 0 |
| claims_conservation | 7 | 0 | 0 |
| fractional_atomic | 13 | 0 | 0 |
| permissioned_burn_wall | 2 | 0 | 0 |
| fractional_compaction | 12 | 0 | 0 |
| escrow_pda_handover | 2 | 0 | 0 |
All 43 tests executed. None was skipped or never ran. Every target has its separate test-<target>.log in the run root. Nine hermetic runner controls also passed locally through python3 -m unittest discover -s tools/gates/tests -p test_fractional_runner.py; these controls test orchestration, not the protocol.
The red control and accepted poststates
At intake source 0a869a76633f2755ef3d359c2d8b63b5b6447f3b, all five transaction cases in claims_conservation failed at CoreState::encode with InvalidPhase; its native census control was the sole passing case. The fixture changed Phase::Open while leaving readiness Prepaid. Core's existing state invariant requires Consumed. This was a fixture defect, not a reason to loosen Core's codec. Baseline: test-claims_conservation.log in the evidence root named below. The two previously unrun founder-bond refusals executed as part of the seven passing founding tests.
seed_open_market now labels the installed prestate explicitly. Claims founding executes the real ELF and creates the aggregate and Positions; Core Open itself is seeded here and earns no execution credit. The refunding accepted round trip checks backing, ordinary liabilities, failure escrow, allocation, surplus donations, balances, and restoration. The additional categorical round trip checks every coordinate including failure, the aggregate supply and exact owner and vault token restoration, and proves no refunding escrow was created.
Every conservation hostile still asserts its exact refusal enum. It now also compares complete named account pre/poststates, including the earlier token approval in the same transaction; only the payer's exact bank-quoted fee may change. A refusal reaching the wrong conjunct therefore cannot satisfy the campaign by merely being an error.
Transaction observations
Evidence root: hbox:/tank/dregg-build/dclutch-codex-evidence-20260907. committed-evidence/ contains 77 instrumented transactions across 47 labels from founding, conservation and compaction. The other three passing targets are not claimed to emit census observations. claims-committed-folded.json was folded once from that directory; the census admitted 148 observations into claims-committed-ledger.json using the committed bindings. Bindings were written from observed transactions; a token donation contributes no protocol route, and seeded Core Open is not credited.
| Evidence artifact | SHA-256 |
|---|---|
| claims-committed-folded.json | a86af15e6612e88d0bceecadbff18ccae95525e4907a2ddb9bc5dd8af6ef9c3b |
| claims-committed-ledger.json | 035aa721cc17be17293130cfe67d6cb3664e82a880bc418a50e38596daa19051 |
| committed-runner/run.e0CwMY/results.tsv | f5f9f871a62ad627711431439f2dcb9ffd74a5ebd266d201476b326a4b4d517e |
Executed ELF identities
All hashes are SHA-256. These are the exact files in the run's sbf-out/; elf-sha256.txt is retained beside the logs.
| ELF | SHA-256 |
|---|---|
| dclutch_claim_check_escrow_signer_test_sbf.so | 237c63a884aafc80b8a22f6f6b5bf3d4bc1d3c9b6156270e2fa05fadad9c38b1 |
| dclutch_claims_founding_test_caller_sbf.so | 3874f94b7d4e269fb8e158a306c67839c2ab96edabec6763917517a1698969aa |
| dclutch_claims_sbf.so | 19a3be2fcb625bf56166f8b638adba2e7d970421d1a08e9fdd1bc63d7ddaef5d |
| dclutch_core_sbf.so | 7013924151b5fbd1f3d5abf0c88fb6199adaf2a6ba7f2a91da89014c4cf086a9 |
| dclutch_custody_sbf.so | 178bb636165036ef061dc2268868c845fed47e3f9bdd1632d2c06161b242a5b7 |
| dclutch_fractional_atomic_test_caller_sbf.so | d485ac634df30416a23cc25eb1dbf068bd3cbc7e395f2f4e8e3b54ca0d4d2335 |
| dclutch_fractional_compaction_test_caller_sbf.so | 76d26ce2d2e28ae4cc4d7bcc7a2237f6b75f828036f0dfad7702c8ff9c54a62c |
| dclutch_registry_sbf.so | 8eb3ccc0e9d0f895521be92b48f5ce6ac912fdca17c79018148581056150fc54 |
| dclutch_rent_sbf.so | 100f211918acc5764fd797b4e2070bc0dd7b1b6ce095351c2494ac6355653143 |
| spl_token_2022.so | e2acdfb750881462ad613a15cc9c54ae17ce066580e867e1e635fbdfe01f5697 |
Reference convergence and remaining scope
tools/gate reference --converge ran from the clean detached worktree hbox:/tank/dregg-build/dclutch-codex-reference-20260907 at the accepted source and reached its fixpoint on pass two (reference-cac1.log). The generated reference discharges the conservation wiring block. It also corrected existing General ABI drift: obsolete entries in generalSuccessorV5 disappeared and current generalClearingPriceV1 and generalOrderV2 pages appeared. No generated page was edited by hand.
This batch closes the initial unrun Claims cases and repairs the runner's accounting. It does not close any unexecuted family, the full local-validator lifecycle obligation, the checked release, or the fresh devnet/site/simulator request. Those remain in the active development-wave plan.