dind #21 (draft): daemon scaffold + preferences/capabilities (P4-1.1, P4-2.1/2.2) - #32
Draft
Santiagocetran wants to merge 40 commits into
Draft
Conversation
Documentation/ now describes only what exists in code on develop (not the live Sepolia deployment), split by audience: - public/ — participant-facing: setup, cli-reference (was common.md), getting-started, manifest, services, roles/, workflows/, guides/ - technical/ — internal: contracts/ (per-contract refs), mechanisms/, services/, testing/, upgradable-contracts/, ARCHITECTURE.md (placeholder moved from Developer/, lands via P3-DOC1) Developer/ holds forward-looking material: - design/ (new) — MECHANISM_DESIGN.md, suggested-staking-mechanism.md (moved from Documentation/technical/mechanisms — it specifies planned behavior, not current), feasibility-report.md - proposals/ (was tooling/) — proposals for tools not yet built - issues/, tasks/, discussion/, rejected-ideas/ unchanged Also: - Rewrote both index READMEs with the placement rule and layout - Added "where does a new doc go" + graduation rule (shipped designs get rewritten as current-state in Documentation/technical/, not moved) - Fixed all repo-internal links to moved files (CLAUDE.md, README.md, Developer/*, dincli/docker/node/README.md) and repaired relative links inside moved files; fixed dead blob/main GitHub URLs to blob/develop - Added DevNet 2.0 mechanism design doc (design/MECHANISM_DESIGN.md) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…nnect-wallet fixes + OWS spike Squash of three commits from the PR InfiniteZeroFoundation#16 review branch (634ce1e, 5cd1a9b, 6905808): - list-accounts source tag: [..] was parsed by Rich as a markup tag and dropped; use (..) so imported/created/legacy renders. Docker node README: .session is a file, not a directory (review bugs InfiniteZeroFoundation#1, InfiniteZeroFoundation#6). - connect-wallet: confirm before overwriting an existing named keystore (--yes/-y to bypass, mirrors send-eth), never encrypt a new keystore with a stale in-memory cached password, and parse .env once per unlock instead of per _get_password call (review items InfiniteZeroFoundation#2, InfiniteZeroFoundation#3, InfiniteZeroFoundation#5). Adds 7 tests. - OWS delegation feasibility spike (Developer/discussion/ ows-delegation-feasibility.md): hands-on verification that OWS signs an EVM tx without exposing the key; scoped-policy delegation remains unproven/follow-up. Corrects wallet-setup.md's broken OWS export-keystore flow. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Cherry-picked from PR InfiniteZeroFoundation#28, relocated from Documentation/public/guides/ to Developer/proposals/: this documents research into a storage provider dincli does not yet integrate (no FOC/filecoin-pin adapter exists in ipfs.py), so per Developer/README.md's placement rule it belongs with forward-looking tooling material, not with docs of what already exists on develop. Replaces the Storacha guide (PR InfiniteZeroFoundation#26, closed) as the storage recommendation for Discussion InfiniteZeroFoundation#18: Storacha's infrastructure (console, API host, and both storacha.link/w3s.link retrieval gateways) is down as of this research -- Cloudflare Error 1000 on the apex domain, NXDOMAIN on console/API subdomains, gateways redirecting into the broken root. Not a frontend-only issue. Covers Filecoin Pin (the filecoin-pin CLI/library, built on FOC's Synapse SDK) in the same three-part structure as the Storacha guide: Web setup, CLI upload/retrieve, and authentication -- wallet keys and the Session Key delegation model, which turns out to be a much closer parallel to Storacha's UCAN delegation than raw-key-or-nothing. Fact-checked against the filecoin-pin GitHub repo directly (README, CLI source in src/commands/, documentation/retrieval.md and content-routing-faq.md) rather than docs.filecoin.io, which 404'd on the CLI walkthrough page during this research. Caught and corrected one safety-critical error from an earlier research pass: the CLI defaults to Mainnet (real funds), not Calibration testnet as initially assumed.
Discussion InfiniteZeroFoundation#18 hands-on testing ruled out both providers this doc recommended: Lighthouse retrieval is payment-gated (fails the "readers pay nothing" requirement), and Storacha's infrastructure is down/deprecated. Point at the new Filecoin Onchain Cloud proposal (Developer/proposals/filecoin-onchain-cloud.md) instead of leaving a stale recommendation for the next reader to walk into. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…n guide (relocated) Manually merged PR InfiniteZeroFoundation#28 by Similoluwa Abidoye. Content unchanged from the PR except relocation from Documentation/public/guides/ to Developer/proposals/: FOC/filecoin-pin is not integrated into dincli (no code changed here, ipfs.py still only supports env/filebase/custom), so per Developer/README.md's placement rule this is forward-looking research material, not documentation of what exists on develop. Also updates Developer/discussion/add-filecoin-support.md, which previously recommended Lighthouse and Storacha as the best Filecoin-backed providers -- both since ruled out by this same research (Discussion InfiniteZeroFoundation#18) -- so it no longer contradicts the new proposal doc. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
InfiniteZeroFoundation#20) v0.3 proposal for the dincli/sdk/ extraction: five design principles, the JSON envelope contract, DinError taxonomy with reserved tx/rpc subcodes, non-interactive DinSession/SignerProvider, plan/apply (three-phase for CID-producing ops) confirmation model, richer TxReceiptInfo, and backward- compat guarantees. Candidate seed for P3-DOC7. Incorporates two rounds of code-audit against real call sites. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…niteZeroFoundation#22) Manually merges the Part 1 deliverable of task_060726_4 from Similoluwa Abidoye (@Abidoyesimze, PR InfiniteZeroFoundation#22, branch security/foundry-src-review). The 7 contracts in foundry/src/ are byte-identical between the audit's pinned commit (d136ff3) and current develop, so the findings and line references apply as-is to the live contracts. - Documentation/technical/audits/foundry-src-security-review.md (relocated from the PR's Developer/audits/ path) - foundry/test/SecurityFindings.t.sol — 9 PoC/verification tests, all passing (forge test --match-contract SecurityFindingsTest). Findings: 0 Critical, 4 High, 4 Medium, 8 Low/Informational. Open follow-ups tracked on PR InfiniteZeroFoundation#22: severity of H-1/H-2 (candidate for Critical) and a measured spec-scale gas number for H-1. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…h real forge measurements Follow-up to PR InfiniteZeroFoundation#22 review feedback: 1. Severity: H-1 (Sybil-cheap gas-DoS) and H-2 (zero-CID sentinel collision) elevated to Critical (now C-1/C-2). Both permanently brick a Global Iteration on non-upgradeable contracts with no recovery path, and both are cheap (C-1) or free (C-2) for a single unprivileged participant to trigger. Remaining High findings renumbered H-1/H-2; no finding content changed, only labels. 2. Gas number: added test_gas_finalizeEvaluation_and_slashAuditors_atScale, measuring real forge gas at two registrant scales (10 and 30 batches at demo params) and deriving the real marginal per-batch and per-(auditor,model)-pair cost from the slope between them. Params.auditorsPerBatch/modelsPerBatch has no setter, so literal spec-scale batch *internals* can't be deployed without modifying the contract (out of scope) -- the report is explicit that the final spec-scale projection layers a structural extrapolation on top of real measured data, not a second measurement, and that the slashAuditors number is a measured floor (its early-break behavior under an all-Sybil, zero-vote attack undermeasures the function's true worst case). Result: 223.5M gas (finalizeEvaluation) / 169.5M gas floor (slashAuditors) projected at full spec scale -- 5.6x-7.5x over Optimism's ~30M block limit, corroborating and quantifying the original estimate rather than changing its conclusion. 10/10 tests passing in SecurityFindingsTest. Reproduced the maintainer's UpgradeValidation.t.sol FFI/npx failure independently in a second environment -- confirmed not related to this report's PoCs.
…ary test (InfiniteZeroFoundation#20) First clean-seed extraction for the SDK layer: - dincli/sdk/cid.py: moved verbatim from services/cid_utils (zero cli deps); services/cid_utils.py kept as a back-compat re-export shim (13 importers). - dincli/sdk/errors.py: DinError hierarchy + reserved tx/rpc subcodes + allowlist-oriented sanitize_details (proposal §4/§4a). - tests/test_sdk_boundary.py: fresh-subprocess import-boundary test enforcing 'sdk imports no typer/rich/dincli.cli', plus error/sanitizer unit tests. runtime/ipfs/contracts deferred to the loaders step: they import cli.utils/log at module load, so moving them now would fail the boundary test until config/manifest/web3/log are extracted. Unit suite green (76 passed, 1 skipped). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…iteZeroFoundation#20) Move the dependency-root loader cluster out of dincli/cli into dincli/sdk, leaving the old paths as re-export shims: - sdk/config.py: config/env/network/IPFS-config resolution + constants + IPFSConfig; console.print -> logger; logging.getLogger("dincli") directly (no sdk.log import -> no config<->log cycle). - sdk/log.py: logger + quiet _read_log_level() bootstrap (no missing-config warning at import). - sdk/web3.py: get_w3 (ConnectionError kept; # TODO(sdk-keystones)). - cli/utils.py: re-exports placed above the wallet constants they feed; wallet/keystore/manifest/tx code stays. - cli/log.py: re-exports logger + stdlib logging (dincli.cli.context needs both). Tests: add test_sdk_config.py (17) + test_sdk_web3.py (2); migrate test_connect_wallet's config-file patch to sdk.config. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…n-seed (InfiniteZeroFoundation#20) - sdk/ipfs.py: 3-provider upload/retrieve from services/ipfs; imports redirected to sdk.config/sdk.log/sdk.cid; progress prints -> logger; # TODO(sdk-keystones): IpfsError. - sdk/contracts.py: ABIs + get_contract_instance; lazy get_w3 from sdk.web3. - services/ipfs.py, cli/contract_utils.py: re-export shims. - tests/test_ipfs_config.py: migrate config-file + ipfs-global patches to sdk.config / sdk.ipfs (re-export does not redirect a moved fn's globals). The dincli/sdk package now holds a dependency-clean foundation (cid, errors, config, log, web3, ipfs, contracts); the import-boundary test enforces no typer/rich/dincli.cli imports across all of it. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Target architecture for the full stack — DAO/platform/task contracts, indexer, SDK, dincli/dind, IPFS layer, din-node/worker — with layered Mermaid + ASCII diagrams, key flows, and P4 dependency order. Links the live diagram on mermaidviewer.com and indexes the doc in Developer/README. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…SON logging (P4-1.1) (InfiniteZeroFoundation#21) New dincli/dind/ package (SDK-independent P4-1.1 framework; no role automation): - CLI lifecycle: dind start | stop | status, all honoring --state-dir / DIN_DIND_STATE_DIR via a shared resolver (config.py); start refuses a live PID (exit 1), status is PID-based and enriched with /health when reachable. - Event loop (daemon.py) with injectable stop_event/max_ticks; heartbeat is a daemon_meta update (not a queue row). SQLite (WAL) job store (state.py): jobs + daemon_meta, enqueue/claim/complete/fail/checkpoint, retention from day one. Graceful SIGTERM/SIGINT (signals.py): running -> pending checkpoint, clean shutdown, PID removal. - HTTP /health (health.py, stdlib ThreadingHTTPServer) with resource summary; binds 127.0.0.1:8787 by default (DIN_DIND_HEALTH_HOST/PORT), records the bound address in daemon_meta. - Structured JSON logging (logging.py): JsonFormatter + idempotent configure_logging() attached to the "dincli" logger with propagate=False and an explicit INFO-default level (via _resolve_log_level, mirroring sdk/log.py) so daemon INFO lifecycle logs emit regardless of import order. - systemd/launchd unit examples (examples/); pyproject: dind console script + package-data for examples. No dincli/sdk or dincli/cli files modified — purely additive, keeps feat/din-daemon a clean superset of feat/din-sdk. Tests: 40 dind tests (config/process/boundary/state/daemon/health/logging), full suite 137 passed / 1 skipped; end-to-end start/health/stop verified. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…niteZeroFoundation#21) Wire the din-node container to run dind as its main process now that the daemon exists (the pre-marked handoff points): - Dockerfile: CMD -> ["dind", "start"] (was sleep infinity). - docker-compose.yml: command -> dind start; add a healthcheck hitting /health so Docker auto-restarts an unhealthy daemon (T0.2c). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…proxies (PR InfiniteZeroFoundation#13) Land the platform-upgradeable work from PR InfiniteZeroFoundation#13: - DinCoordinator, DinToken, DinValidatorStake, DINModelRegistry converted to initializer-based transparent-proxy implementations - V2 stub implementations under hardhat/contracts/upgrade/ with upgrade test suites under hardhat/test/ - deploy-platform.ts / upgrade-platform.ts scripts plus shared deploy/{constants,helpers,types}.ts and test/helpers/platform.ts - MockSlasher test mock, hardhat-upgrades tooling deps, updated DinCoordinator/DinToken ABIs - .gitignore: stop ignoring hardhat/scripts and hardhat/test so the deploy/upgrade tooling is tracked on develop Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…) with dincli harness, docs, wallet split Follow-up integration on top of the PR InfiniteZeroFoundation#13 contract work: - dincli test harness (tests/dincli/) updated to deploy the platform via the interim deploy-platform.ts script + import-deployments flow; hardhat/deployments/localhost.json and din_info.json tracked for it - dincli: register-wallet/connect-wallet split follow-ups in cli/{context,core,system,utils}.py and main.py; regenerated DINModelRegistry/DinValidatorStake ABIs - Documentation/technical/contracts/hardhat/: per-file docs for the new deploy/upgrade tooling, V2 contracts, and test suites (with coverage gaps); refreshed platform contract docs for the proxy conversion - Documentation/technical/testing/: new docs for dintoken, ipfs-config, cache-client-dp, connect-wallet test suites; testing-guide and containerization-guide updates - Documentation/public/: wallet-setup, keystore-migration, setup, getting-started, cli-reference, model-workflow updated for the wallet split and upgradeable deployment - Developer/issues/dincli-native-proxy-deployment.md: backlog Option C (native dincli proxy deployment); retire technical/ARCHITECTURE.md Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
… the loader tier (InfiniteZeroFoundation#20) Move the manifest / din_info functions and ServiceRuntimeContext into dincli/sdk, leaving the old paths as backward-compat shims: - sdk/manifest.py: load_din_info, save_din_info, load_cid_services, download_manifest, get_model_info, get_manifest, get_manifest_path, get_manifest_key, is_ethereum_address; patchable DIN_INFO_PATH / CID_SERVICES_PATH constants; logging.getLogger("dincli") directly (no sdk.log import -> no cycle). Presentation stripped: cache_manifest's negative-id guard -> ValidationError; the info-block printing lifted out into a pure get_model_info returning data; the freshness warning -> logger.warning. - sdk/runtime.py: ServiceRuntimeContext + build_service_runtime_context, import repointed to sdk.manifest. - cli/utils.py: re-exports (above the wrapper); cache_manifest is now a thin CLI wrapper composing download_manifest + get_model_info + console.print and mapping ValidationError -> typer.Exit(1) (main.py has no DinError handler), reproducing the original info output verbatim. - services/runtime.py: re-export shim. Completes the SDK loader tier (cid, errors, config, log, web3, ipfs, contracts, manifest, runtime). No sdk<->cli cycle; import-boundary test now covers manifest+runtime. Tests: +22 (test_sdk_manifest), suite 117 passed / 1 skipped; behavior parity + tests/dincli import integrity verified. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…/home/azureuser constants.py now resolves everything from the repo checkout, home directory, or environment: repo root via __file__, venvs via ~/my_venvs/<name>, npx via the newest ~/.nvm/versions/node/*/bin/npx, scratch dir ~/tempdir/dincli. All resolutions are overridable via env vars (PYDIN_PYTHON, TORCHENV_PYTHON, TORCHENV_SITE_PACKAGES, NPX_BIN, IPFS_BIN, DIN_TEST_TMPDIR), read from the real environment first and then the repo-root .env. conftest.py and test_04_gi.py consume the shared constants instead of their own literals. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…ences Rewrite all 101 remaining /home/azureuser occurrences across 17 files in Documentation/ and Developer/: - markdown links to repo files become relative links (GitHub-navigable) - shell/repo paths use the /path/to/devnet placeholder - venv and scratch paths use ~/my_venvs/... and ~/tempdir/dincli/ - HR profile pointers in migrate_to_foundy.md use ~/projects/HR/... - DIN_STATE_DIR example uses /home/<user>/.din-node; drop a stray duplicate mkdir line in containerization-guide.md Also repoint references to the six per-model service files (client/aggregator/auditor/model/scoring/modelowner.py) from dincli/services/ to cache_model_0/services/ — they moved there when services became per-model — and retarget the proposed contribution.py module to cache_model_0/services/ for consistency. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Working summary of Abraham Nash's DIN white paper (Jun 2026) with a gap/alignment checklist mapping paper mechanisms to DevNet status, feeding the P3 mechanism-design push. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…acking into docs - design/p3-design-plan.md: coordination page for the P3 mechanism-design push — mechanism -> design sources -> GitHub issue map (issues 36-43), open-decision -> discussion map (discussions 44-46), white-paper gap routing, definition of done - ROADMAP.md: P3 status note (delayed pending designs; P4 active in parallel; dates not binding) - MECHANISM_DESIGN.md, README.md: cross-references Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
… (PR InfiniteZeroFoundation#35) Original work by Robert (@robertocarlous, PR InfiniteZeroFoundation#35, branch feature/foundry-platform-deployments, head 45dcda1), implementing issue InfiniteZeroFoundation#34: DeployPlatform.s.sol, UpgradePlatform.s.sol, four V2 upgrade stubs (DinTokenV2, DinCoordinatorV2, DinValidatorStakeV2, DINModelRegistryV2), the forge test suite (DeployPlatform.t.sol), dincli's `system import-deployments --foundry` flag, and the foundry.toml/package.json/.gitignore changes needed to run them. Three blocking bugs found in local verification, fixed here: - DeployPlatform.s.sol: vm.writeJson has no parent dir on a fresh clone (foundry/deployments/ is gitignored and not committed) — call vm.createDir before vm.writeJson. - All four V2 stubs: a plain `//` comment sat between the two `///` NatSpec lines, splitting the doc block so `@custom:oz-upgrades-from` never reached solc — UpgradePlatform.s.sol failed upgrades-core validation for every contract. Moved the `//` explanation above the NatSpec block so the `///` lines stay contiguous. - Both scripts' own documented "Usage (from repo root)" command didn't compile (foundry.toml/remappings.txt live in foundry/, not repo root) — changed the usage comments to `cd foundry && forge script ...`. Verified end-to-end: forge test (40/40 passing), and both scripts run against a local chain (deploy -> deployments JSON -> dincli system import-deployments --foundry -> upgrade, version() == 2 through the proxy for all four contracts). Co-Authored-By: umeradl <umermajeed.cto@gmail.com> Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
… harness tests/dincli/ can now deploy the platform contracts via either the new forge script from PR InfiniteZeroFoundation#35 or the existing hardhat script, selected by PLATFORM_DEPLOY_TOOLCHAIN (tests/dincli/constants.py; default "foundry", matching `dincli system import-deployments`'s own default). Overridable via env var / .env for anyone who wants to keep exercising the hardhat path. - constants.py: FOUNDRY_DIR, FORGE_BIN, PLATFORM_DEPLOY_TOOLCHAIN, HARDHAT_DEV_ACCOUNT_0; DEPLOYMENTS_FILE now resolves under whichever toolchain's deployments/ dir. - test_01_platform.py: test_deploy_platform_via_script branches on the toolchain (forge script vs hardhat run); test_import_deployments_into_din_info keeps passing an explicit --file since import-deployments' own --foundry/--hardhat default-path resolution is relative to cwd, and the harness's cwd is the isolated DIN_TEMP scratch dir, not the repo root. - conftest.py: managed_services now starts the matching chain backend (Anvil via foundry/anvil.sh for "foundry", the existing Hardhat node for "hardhat" — both on chain-id 1337) and runs `forge build` when the foundry toolchain is active; `npx hardhat compile` still always runs since task-level contract deploys and dump-abi tests need hardhat's ABIs/bytecode regardless of which toolchain deployed the platform. Verified: full tests/dincli/test_01_platform.py (8/8) under both toolchains, and test_01 through test_03 (35/35) end-to-end under the new foundry default. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Two follow-up fixes to files landed in b40fe2e (PR InfiniteZeroFoundation#35 foundry platform deploy/upgrade parity): - foundry/test/DeployPlatform.t.sol: the DinCoordinatorV2/ DinValidatorStakeV2/DINModelRegistryV2 imports sat mid-file, after the MockTaskContract definition — moved them up next to the DinTokenV2 import at the top, with the rest of the imports. - foundry/script/DeployPlatform.s.sol: added a `forge clean` step to the documented "Usage (from repo root)" comment, right after starting anvil. Without it, deploying then testing back-to-back after a recompile can hit upgrades-core's stale build-info bug ("Found multiple contracts with name ...") — reproduced again while re-verifying this change; `forge clean` resolves it. Also ran `forge fmt` over both files (line-wrapping only, no logic changes). Re-verified: forge clean && forge test — 40/40 passing. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Adds staking-design.md and tokenomics-design.md, and updates the mechanism design, staking mechanism, and P3 plan docs to reflect them. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…ions Records outcomes from Abraham's Slack sync (2026-07-21) against the open decisions tracked in MECHANISM_DESIGN §9, staking-design.md §5, and tokenomics-design.md §6: slashed-stake destination (50/50 burn/treasury), depositAndMint cap/retirement, emission decay with MAX_SUPPLY left open, dual-source burn policy, delegation/tombstoning scope, and the white-paper InfiniteZeroFoundation#46 alignment items. Fee denomination is resolved as a DIN/ETH split (protocol fees in DIN, validator network fees in ETH), not the DIN-only recommendation these docs previously carried. GitHub Discussions InfiniteZeroFoundation#44/InfiniteZeroFoundation#45/InfiniteZeroFoundation#46 themselves are not yet updated to reflect this — noted in p3-design-plan.md so the "resolved" checkbox isn't marked done prematurely. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…iteZeroFoundation#37/InfiniteZeroFoundation#42/InfiniteZeroFoundation#43) Scopes a 5-day platform-contract task combining the now-resolved staking, tokenomics, and fees/treasury designs: new DinTreasury and DinFeeRouter contracts, DinToken.burn(), DinCoordinator mint-cap/faucet-retirement/ treasury-routing, DinValidatorStake governable params + real jailing, and a DIN-denominated fee path alongside DINModelRegistry's existing ETH fees. Includes the exact deployment/wiring order and test plan. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…nfiniteZeroFoundation#39/InfiniteZeroFoundation#40/InfiniteZeroFoundation#41/InfiniteZeroFoundation#38) Adds task_210726_6 (scoring validation, commit-reveal auditing, encrypted test-data distribution, per-GI reward engine bootstrap) plus a prior-work reference doc mapping BlockFlow/2CP/Shapley-value-data-valuation/IPFS onto the task's scoring and auditing mechanisms. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
…capabilities task (issues InfiniteZeroFoundation#20/InfiniteZeroFoundation#21) Co-Authored-By: umeradl <umermajeed.cto@gmail.com>
OZ v5 TransparentUpgradeableProxy deploys one ProxyAdmin per proxy. The previous script read the admin from the token proxy only and wrote a single proxyAdmin key, leaving the other three addresses unrecorded. Now reads getAdminAddress for each of the four proxies and writes proxyAdminToken, proxyAdminCoordinator, proxyAdminStake, and proxyAdminRegistry into foundry/deployments/<network>.json.
Same fix as the Foundry script: read the ProxyAdmin address for each proxy individually and write proxyAdminToken, proxyAdminCoordinator, proxyAdminStake, and proxyAdminRegistry into the deployments JSON. PlatformAddresses type updated to match.
Update _DEPLOYMENTS_TO_DIN_INFO to map the four proxyAdmin* keys (proxyAdminToken, proxyAdminCoordinator, proxyAdminStake, proxyAdminRegistry) to their corresponding proxy_admin_* keys in din_info.json, replacing the single proxyAdmin → proxy_admin mapping.
…proxy-admin-addresses Fix/proxy admin addresses
…severity-and-gas-followups audit(foundry): elevate H-1/H-2 to Critical, add real gas measurements
# Conflicts: # tests/test_connect_wallet.py
… validate predicates (InfiniteZeroFoundation#20) Move GIstateToDes/GIstateToStr/GIstatestrToIndex and their backing data (states, stateDescription, GIstate_to_index) verbatim from cli/utils.py into a pure dincli/sdk/state.py; utils.py now re-exports them as a shim (same pattern as the ipfs/runtime waves) so existing call sites keep working unchanged. Add a GIState IntEnum (0-based, matching the Solidity enum ordinals) and two validate_* predicates lifted from DinContext's GI-state guards (validate_gi_state_equals / validate_gi_state_at_least), raising the SDK's ValidationError instead of printing + typer.Exit. The two context.py guards now delegate the decision to these predicates while keeping identical console output and exit codes (no CLI behavior change). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…eroFoundation#20) Implement to_envelope() + a metadata-driven JSON encoder per proposal §3: the {status, data, error, meta} envelope with schema_version/sdk_version, and the encoder rules — uint256_string -> decimal string, omit, small-int passthrough, bytes/HexBytes -> 0x hex, address -> checksummed, enum -> name, Decimal -> string, and None -> null (except tagged omit). Extend the import-boundary test to assert state and serialize import clean of typer/rich/dincli.cli. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
…2.2) (InfiniteZeroFoundation#21) Part B of task_220726_7 — local-only daemon config + hardware introspection. Preferences (P4-2.1): - dincli/dind/preferences.py: Preferences dataclass (domain, risk_tolerance, min_expected_reward, privacy_constraints) + load/save JSON helpers that return defaults on a missing or empty/whitespace file. Stored at StateDirs.preferences_path (state_dir/preferences.json), reusing the existing path helper. - `dind preferences show` / `set` — partial updates (unset flags leave stored values untouched), risk-tolerance validation. Capabilities (P4-2.2, hardware/local slice): - dincli/dind/capabilities.py: CapabilitySummary + detect_capabilities() — CPU count/speed, RAM via os.sysconf + /proc/meminfo (no new dependency), disk, best-effort GPU, and bare socket reachability probes for RPC/IPFS (endpoints via sdk.config, default https->443/http->80, None when unconfigured). Pure score_capabilities() / compatible_with() for later task-matching. - `dind capabilities` command; both new command groups resolve --state-dir first. - Factor the shared resource_snapshot() (CPU+disk+RAM, fast/local) into /health's resources block — removes duplicated disk logic and keeps GPU/network probes off the frequently-polled path. Tests: test_dind_preferences.py, test_dind_capabilities.py (network-hermetic), test_dind_health.py extended (new fields + a guard that a poll runs no socket probe). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Santiagocetran
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Draft / WIP — progress on #21, not merge-ready. Opened early to keep feature-branch work visible. Base is
feat/din-sdk(the branch this stacks on), notdevelop.What this contains
The P4-1.1 daemon scaffold plus the P4-2.1 / P4-2.2 local slice (
task_220726_7Part B).P4-1.1 — daemon scaffold (
dincli/dind/)SDK-independent framework; no role automation / signing (that waits on the SDK contract):
dind start | stop | status, all honoring--state-dir/DIN_DIND_STATE_DIRvia a shared resolver.startrefuses a live PID (exit 1);statusis PID-based, enriched with/health.stop_event/max_ticks); SQLite (WAL) job store (jobs+daemon_meta) with enqueue/claim/complete/fail/checkpoint + retention; heartbeat is a meta update.running → pendingcheckpoint, clean shutdown, PID removal./health(stdlibThreadingHTTPServer); binds127.0.0.1:8787by default (DIN_DIND_HEALTH_HOST/PORT), records the bound address.JsonFormatter+ idempotentconfigure_logging()on thedinclilogger.systemd/launchdunit examples;dindconsole script; container handoff (din-nodeDockerfileCMD→dind start, composehealthcheckon/health).P4-2.1 / P4-2.2 — preferences + capability detection (Part B, this task)
dincli/dind/preferences.py—Preferencesdataclass (domain,risk_tolerance,min_expected_reward,privacy_constraints) + JSONload/save(defaults on missing/empty file), stored atStateDirs.preferences_path. Commands:dind preferences showanddind preferences set(partial updates — unset flags leave stored values untouched; risk-tolerance validated).dincli/dind/capabilities.py—CapabilitySummary+detect_capabilities(): CPU count/speed, RAM (viaos.sysconf+/proc/meminfo— no new dependency), disk, best-effort GPU (nvidia-smi//dev/nvidia*), and bare socket reachability probes for RPC/IPFS (endpoints resolved viasdk.config; defaulthttps→443/http→80;Nonewhen unconfigured — not a live RPC call). Purescore_capabilities()/compatible_with()for later task-matching. Command:dind capabilities(JSON)./healthresourcesnow uses a sharedresource_snapshot()(CPU+disk+RAM) — removes the duplicated disk logic and deliberately keeps GPU/network probes off the frequently-polled path (the full detect backsdind capabilitiesonly).Verification
preferences/capabilitiestests (network-hermetic; a guard asserts a/healthpoll runs no socket probe) + thedindimport-boundary test (daemon imports nodincli.cli.*).tests/dincli/*, needs Hardhat/Foundry) and the Torch DP test.feat/din-sdk(which carriesdevelop+ the SDK state/serialize work).Scope / boundaries
dincli/sdkanddincli/cliare only consumed, not modified here beyond the sync — purely additive daemon work.wallet.py/session.py/SignerProvider§2,tx.send()§5b): the dedicated next task, flagged for @umeradl's review given the signing sensitivity. Sodindstill does no signing / on-chain submission;/healthRPC-reachable + wallet checks beyond the bare probe stay TODO pending role wiring.dindarchitecture doc) isn't written yet — these choices (SQLite state, threaded loop, stdlib health, JSON logging, local preferences/capabilities) effectively seed it; happy to adjust to spec.