✔ a budget pause persists concrete text/tool/history/image breakdown and never dispatches an oversized request (2997.0554ms) ✖ explicit frozen preview resumes the same Task in a new Run, preserves completed tools and does not change global config or selected model (3513.8586ms) ✔ insufficient budgets, silently clamped capacity and arbitrary config fields reject without a physical request or mutating task budget (3520.4991ms) ✔ changed session, changed registry, expiry and altered frozen preview all require a new explicit preview (3275.0687ms) ✖ preview preserves unknown invocation status and increasing Task allowance is committed only with confirmed new Run (3247.2136ms) ✔ pairing is short lived, single-use, generation-bound and validates exact permissions (965.5924ms) ✔ device transaction failure neither consumes the ticket nor leaves a credential (789.0047ms) ✔ devices have project and action boundaries; rotation, expiration, revocation and host changes reject old authority (1114.822ms) ✔ persistent registry stores hashes and a bounded audit, never raw credentials (1040.811ms) ✔ real TLS verifies IP SAN, CA chain, encrypted restart and public fingerprint (1273.6199ms) ✔ TLS sealing failure leaves listener closed and no plaintext or successful TLS registration (742.8388ms) ✔ HTTP pairing is concurrent single-use; secure cookie and strict origin/host checks (1093.9074ms) ✔ gateway never exposes management routes, arbitrary assets or unpaired project data (1121.7398ms) ✔ authorized project events paginate from original seq without leaking other project events or payloads (2176.9958ms) ✔ remote task projection excludes raw model history, host configuration, tool input and credentials (1064.2894ms) ✔ permission checks apply to direct remote mutation, not only UI controls (1064.1012ms) ✔ approval fingerprint and exact task/revision are bound; duplicate requests never execute twice (977.3428ms) ✔ unknown original operation is recorded and never replayed after response failure (1294.444ms) ✔ revocation and gateway closure reject old access without invoking runtime stop (1194.679ms) ✔ rate limits reject pairing brute force and do not persist candidate secrets (988.1917ms) ✔ real Windows CurrentUser DPAPI persists only a sealed CA and reopens it (2135.293ms) ✔ mobile approval executes the original Runtime command exactly once and rejects the later desktop decision (3246.1573ms) ✔ desktop rejection wins a concurrent mobile request and original command remains unexecuted (2258.763ms) ✔ phone stop waits for the original process exit and prevents later heartbeats and command replay (3645.9035ms) ✔ Run replacement during parent-stop wait is rejected rather than stopping the new Run (937.9841ms) ✔ diff pages preserve source SHA; report reads verify blobs and exclude unrelated artifact kinds (1484.4431ms) ✔ hot upgrade refuses an active gateway and restores closed gateway ownership on failure (923.7663ms) ✔ damaged device permissions or expiry fail closed and never overwrite the original registry (1138.4868ms) ✔ device revocation destroys its actual authenticated TLS connection (1193.174ms) ✔ companion observes only current Task screenshots, verifies original PNG bytes and loses access immediately on revoke (1513.6405ms) ✔ schema1 connections remain readable without inventing observed capabilities or migrating original state (1054.5251ms) ✔ protocol, unsupported capabilities and stream choice persist; client claims cannot forge verified evidence (794.7377ms) ✔ ability evidence belongs to exact connection and credential revision; unknown and unsupported reject required routing (0.2757ms) ✔ invalid declarations and inconsistent stream settings reject before registry, revision and credentials mutate (610.7537ms) ✔ saved engine budgets verify all Chat/Responses route variants and reject a mismatched output allowance (1215.7279ms) ✔ owned CC Switch SQLite imports the selected TOML Responses wire_api and leaves unknown clients untouched (1072.1526ms) ✔ Responses executes one original tool through Runtime, carries encrypted Items, reports raw usage and seals exact wire bytes (3025.5661ms) ✔ mixed-protocol pool keeps both actual variants and changes route on 429 without re-executing completed tools (3517.0097ms) ▶ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool ✔ Reject {"body":{"id":"resp_owned","status":"incomplete","output":[{"type":"function_call","id":"fc_partial" (2101.2162ms) ✔ Reject {"body":{"id":"resp_owned","status":"failed","output":[{"type":"function_call","id":"fc_failed","cal (2247.4826ms) ✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"function_call","id":"fc_bad","cal (2081.4415ms) ✔ Reject {"body":{"id":"resp_owned","status":"completed","output":[{"type":"computer_call","id":"not-supporte (1807.2339ms) ✔ Responses incomplete, failed, missing call identity and unsupported hosted outputs never dispatch a tool (8238.5498ms) ▶ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay ✔ chat-completions (1856.2509ms) ✔ responses (1984.9552ms) ✔ invalid JSON in a nonstream Chat or Responses tool group prevents every tool and pool replay (3841.6985ms) ✔ wire adapter preserves tool IDs but never passes Responses-private Items to Chat or another endpoint (0.1784ms) ✔ connection JSON import accepts explicit Responses protocol and continues to reject unknown protocols without reading client sessions (698.2482ms) ✔ HTTP download preserves original malformed SSE bytes and artifact SHA instead of decoding invalid UTF8 (2671.5639ms) ✔ real Chat SSE carries UTF8 and split tool JSON; Runtime executes only after DONE, stores partial progress and original usage (3549.964ms) ✔ Responses typed SSE validates original item identity, terminal data, raw usage and tool result pair (2055.8266ms) ✔ incomplete network stream retains unknown attempt and never automatically switches pool or dispatches fragment tools (1591.9915ms) ✔ stopping an active HTTP stream closes only original request and keeps partial text without executing tools (1387.9026ms) ✔ explicit streaming probe uses actual SSE despite unknown declaration and binds its own original attempt (634.4211ms) ✔ coalesced initial SSE events publish the latest partial state even when the upstream stays idle (1562.436ms) ✔ stream timeout closes the original HTTP response, preserves redacted capture and never retries (1006.5371ms) ✔ SSE accepts CR-only final event boundaries and retains exact invalid UTF8 bytes without dispatch (1.5802ms) ▶ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data ✔ reject 0 (0.4007ms) ✔ reject 1 (0.2435ms) ✔ reject 2 (0.2688ms) ✔ reject 3 (5.7145ms) ✔ reject 4 (0.1532ms) ✔ reject 5 (0.1481ms) ✔ stream parser rejects truncation, finish length, malformed JSON, duplicate sequence, changed final item and oversized data (7.7568ms) ℹ tests 64 ℹ suites 0 ℹ pass 62 ℹ fail 2 ℹ cancelled 0 ℹ skipped 0 ℹ todo 0 ℹ duration_ms 50232.4162 ✖ failing tests: test at tests\budget-resume.test.mjs:27:1 ✖ explicit frozen preview resumes the same Task in a new Run, preserves completed tools and does not change global config or selected model (3513.8586ms) Error: 原任务、连接或预算预览已变化,请重新准备后确认。 at fail (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/core/context/budget-resume.mjs:6:48) at BudgetResume.verify (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/core/context/budget-resume.mjs:57:198) at Runtime.resumeBudget (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/server/runtime.mjs:126:89) at TestContext. (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/budget-resume.test.mjs:30:13) at async Test.run (node:internal/test_runner/test:1409:7) at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) { status: 409 } test at tests\budget-resume.test.mjs:56:1 ✖ preview preserves unknown invocation status and increasing Task allowance is committed only with confirmed new Run (3247.2136ms) Error: 原任务、连接或预算预览已变化,请重新准备后确认。 at fail (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/core/context/budget-resume.mjs:6:48) at BudgetResume.verify (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/core/context/budget-resume.mjs:57:198) at Runtime.resumeBudget (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/src/server/runtime.mjs:126:89) at TestContext. (file:///D:/Ai-project/REA%E9%80%86%E5%90%91/Agent-Workbench/tests/budget-resume.test.mjs:61:13) at async Test.run (node:internal/test_runner/test:1409:7) at async Test.processPendingSubtests (node:internal/test_runner/test:974:7) { status: 409 }