Replace legacy HyperClaw with runtime v2 and post-M6 validation - #9
Open
zhang8128 wants to merge 80 commits into
Open
Replace legacy HyperClaw with runtime v2 and post-M6 validation#9zhang8128 wants to merge 80 commits into
zhang8128 wants to merge 80 commits into
Conversation
Add scoped workspace operations, exact approvals and grants, persisted model checkpoints, verified artifacts, and owned Docker cancellation/reconciliation. Restore complete tool and image transport with bounded execution and public operator controls. Include offline, live Docker/Qwen, and Linux verification gates.
Document shipped tool and operator behavior, mark only M2 complete, and preserve exact macOS/Linux, Docker, live Qwen, coverage, wheel, and review evidence. Record the corrected Linux filesystem harness and remaining milestone boundaries.
Retain the intentional deletion of the legacy CLI/TUI files modified by three newer main-branch provider fixes. Runtime v2 replaces those entry points and has no legacy provider compatibility contract. The merged tree is identical to 94a2fa2. Fresh offline validation passed 587 cases with the expected base-only missing-SDK skip.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
HyperClaw now runs as one installable local package with one daemon owning SQLite state, model requests and tool processes. This branch replaces the legacy multi-provider platform with authenticated HTTP/CLI interfaces, durable session/run/event recovery, exact tool approvals and receipts, scheduling, scoped memory, reviewed skills, admitted documentation MCP, a local web client and an optional durable Telegram adapter.
The post-M6 work adds reproducible installed-wheel acceptance and a one-root walkthrough, mixed HTTP/Telegram-fixture/scheduled recovery tests, populated-state migration/restoration and storage-failure tests, operator accessibility checks, sustained measurements and frozen answer-quality evaluation. The only production change in that testing phase adds run-status accessibility semantics.
Validation: the recorded final matrix reconciles all 644 collected IDs, including 66 new test functions. macOS Python 3.11/3.13 and Linux Python 3.13 each passed 587 offline cases plus the expected base-only skip; Docker passed 22 per platform, Chrome 26, fresh base/MCP wheels 183 public/browser executions with zero skips and 24 matching packaged files, and existing Qwen seven. The 900/3600-second workloads completed 2,700/10,800 runs with zero pending or duplicate effects. All 36 frozen documentation answers and failed attempts are retained. The fresh integration-tree offline suite also passed 587 cases with that same expected skip (152.12 seconds), with exact JUnit evidence retained locally.
Known limits remain explicit: replacing workspace identities during restore hides prior scoped memory; the documentation answer/citation baseline failed despite successful transport and zero unauthorized effects; manual VoiceOver/Safari/human prose inspection remain blocked; real Telegram messaging has not been authorized or tested. Hosted CI was unexecuted at the report cutoff. The report also retains unexplained initial-container disappearance and intermittent Docker lookup observations without attributing their cause.
Integration preserves the deliberate deletion of
cli/hyperclaw.pyandhyperclaw/tui.pywhen incorporating the three newer main-branch fixes to those obsolete files. The resulting tracked tree is byte-for-byte identical to the tested report commit94a2fa2; no runtime compatibility layer or provider fallback is reintroduced.Evidence: testing report, machine-readable results, and runtime design.