Files
Agent-Workbench/docs/release-efficiency-1.23-evidence/local/release-efficiency-related-final-dev.tap
T
2026-11-02 11:58:42 +08:00

60 lines
6.0 KiB
Plaintext
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
✔ fixed browsing contracts reject local host APIs, credential URLs, protocols, cross-origin and arbitrary execution (0.9647ms)
✔ attested desktop queue preserves same-tab observations, original PNG bytes and approval-bound one-shot actions (4043.1042ms)
✔ page data cannot grant domains, interaction or other Task scope; nonce and ownership remain fixed (1618.7814ms)
✔ unknown dispatched submission blocks even after a new observation, rejects late success and retains explicit acknowledgement (3220.6324ms)
✔ downloads require exact approved address, remain original binary artifacts, and cannot reuse consumed observation (1242.3133ms)
✔ revoking shared persistent profile closes all its tabs and clears only that profile after original desktop acknowledgement (1297.31ms)
✔ service recovery never reopens old tabs or replays dispatched actions and keeps unknown original receipts (1088.2607ms)
✔ agent browser flow waits for original Task/Run approval, dispatches once, and returns paired original results (3015.1003ms)
✔ changed approved input and stopped dispatched agent action cannot be accepted or replayed (1205.8232ms)
✔ browser inventory gives bounded metadata and original references instead of repeating full website histories (1104.5051ms)
✔ receipt storage failure settles the original dispatched action as unknown and never hangs or accepts late success (911.299ms)
✔ observe-only device authority cannot read, configure or clear project parsing caches (654.7346ms)
✔ pairing is short lived, single-use, generation-bound and validates exact permissions (849.4866ms)
✔ device transaction failure neither consumes the ticket nor leaves a credential (1117.173ms)
✔ devices have project and action boundaries; rotation, expiration, revocation and host changes reject old authority (1322.5004ms)
✔ persistent registry stores hashes and a bounded audit, never raw credentials (732.5ms)
✔ real TLS verifies IP SAN, CA chain, encrypted restart and public fingerprint (849.745ms)
✔ TLS sealing failure leaves listener closed and no plaintext or successful TLS registration (799.894ms)
✔ HTTP pairing is concurrent single-use; secure cookie and strict origin/host checks (699.474ms)
✔ gateway never exposes management routes, arbitrary assets or unpaired project data (720.7747ms)
✔ authorized project events paginate from original seq without leaking other project events or payloads (1351.413ms)
✔ remote task projection excludes raw model history, host configuration, tool input and credentials (1026.6874ms)
✔ permission checks apply to direct remote mutation, not only UI controls (895.7261ms)
✔ approval fingerprint and exact task/revision are bound; duplicate requests never execute twice (1157.6709ms)
✔ unknown original operation is recorded and never replayed after response failure (1116.387ms)
✔ revocation and gateway closure reject old access without invoking runtime stop (1036.7406ms)
✔ rate limits reject pairing brute force and do not persist candidate secrets (783.0639ms)
✔ real Windows CurrentUser DPAPI persists only a sealed CA and reopens it (3890.4753ms)
✔ mobile approval executes the original Runtime command exactly once and rejects the later desktop decision (2039.9366ms)
✔ desktop rejection wins a concurrent mobile request and original command remains unexecuted (1429.4536ms)
✔ phone stop waits for the original process exit and prevents later heartbeats and command replay (2420.4321ms)
✔ Run replacement during parent-stop wait is rejected rather than stopping the new Run (582.3242ms)
✔ diff pages preserve source SHA; report reads verify blobs and exclude unrelated artifact kinds (1483.6701ms)
✔ hot upgrade refuses an active gateway and restores closed gateway ownership on failure (1008.3032ms)
✔ damaged device permissions or expiry fail closed and never overwrite the original registry (824.0869ms)
✔ device revocation destroys its actual authenticated TLS connection (1324.19ms)
✔ companion observes only current Task screenshots, verifies original PNG bytes and loses access immediately on revoke (825.1659ms)
✔ mobile browser observation is current Task/project scoped, preserves original PNG and exposes no browser control or profile data (791.9388ms)
✔ runtime guides map archived evidence while all original reports and successful/failed source bytes remain independently verifiable (409.5434ms)
✔ missing required research reports fail packaging rather than creating a successful placeholder (485.5317ms)
✔ full verification reports only verified bytes, complete declared scope and actual directory inventory; changed/deleted files still fail (626.4401ms)
✔ preflight and activation continue one monotonic operation bound to installer hash; terminal replay and changed origin are rejected (7067.4513ms)
✔ preflight failures and fixed-candidate mismatches retain the actual reason without publishing a pointer (4958.5477ms)
✔ cancel finishes only the original ready preparation, preserves data and pointer and refuses replay (6635.1659ms)
✔ the installation lock rejects concurrent same-id operations before they can overwrite original progress (3437.0855ms)
✔ original post-publication failure is linked to its rollback receipt and every original byte survives (3842.2689ms)
✔ a candidate mutation at the publication boundary still fails even after an earlier successful full validation (6232.0616ms)
✔ public progress marks a lost process unknown without editing its original record or claiming completion (4996.8052ms)
✔ throwing console/UI progress observers cannot decide the installation result (3660.162ms)
✔ a progress-only write failure after committed publication returns the authoritative receipt and a warning (3546.4625ms)
✔ browser and MCP approvals use actual tool categories; mobile exposes only category while keeping original fingerprint (679.2911ms)
ℹ tests 51
ℹ suites 0
ℹ pass 51
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 97610.3417