Files
Agent-Workbench/docs/reliability-1.25.2-evidence/development/reading-contract.tap
T

52 lines
5.3 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.
✔ unchanged proposals cannot hide repeated original-log reads or identical session inspections (8083.2883ms)
✔ reusing a verified proof never resets identical-state detection or creates a new proof (7473.0495ms)
✔ coverage merges adjacent pages, separates changed sources and never claims old Run coverage in a continuation (0.8894ms)
✔ a mixed turn with new bytes and three repeated modules advances, while repeat-only and independent no-op guards still pause (1.0423ms)
✔ runtime sends the next real HTTP request after a W-shaped mixed-progress turn instead of stopping on its trailing repeats (6647.1744ms)
✔ continuation reads advance through historical pages without a false pause while repeated current-Run pages still pause (1.2804ms)
✔ a new tool page replaces old latest-feedback space and still obeys its escaped protocol budget (0.3703ms)
✔ budgeted artifact pages count escaped protocol bytes and reconstruct original multilingual JSON without losing characters (8.0504ms)
✔ known artifact coverage detects semantic repeats despite changing limits, and advancing pages or direct evidence restore progress (0.307ms)
✔ actual large artifact reads are present in the next bounded HTTP request and next_offset progresses without redoing tools (6622.6739ms)
✔ a real repeated artifact loop pauses the same Task after recorded results, survives reload and only resumes explicitly (8071.6367ms)
✔ actual HTTP continuation after reload sends distinct historically covered pages without false no_progress or replay (9607.0574ms)
✔ continuations distinguish unread source, later original line ranges and already returned result wrappers (644.963ms)
✔ pure-file batch returns unused small-page space while mixed and historical allocation stays conservative (0.2772ms)
✔ observed upstream high input pauses before a new request, preserves reported usage and distinguishes the limiting ledger (748.2003ms)
✔ chat-completions: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (4479.5992ms)
✔ responses: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (4945.74ms)
✔ anthropic-messages: complete actual Unicode/CRLF pages and small source survive the next bounded wire request (4610.5398ms)
✔ operational questions are diagnostic while explicit repair requests remain execution instructions (1.0341ms)
✔ an explicit prohibition on running engineering checks is not interpreted as a requirement to run them (1.7835ms)
✔ file delivery requires a positive instruction, not saved evidence or a prohibited write (0.272ms)
✔ an unsubstantiated text claim cannot complete a requested file delivery (2145.7003ms)
✔ HTML without tests ends at reviewed file delivery and never offers automatic rechecking (3444.4078ms)
✔ diagnostic answers retain the unfinished task, limits and check result, and enforce read-only tools (5056.5468ms)
✔ review continuation is a system action and adds no testing demand or fake user instruction (3822.1035ms)
✔ a provider cannot execute tests against an explicit no-test task even if it ignores advertised tools (2387.7494ms)
✔ a budget pause becomes visible only after its input snapshot can be sealed with the original Run (2386.2671ms)
✔ alternating duplicate file reads and listing files still pauses without treating new pages as duplicates (19015.1185ms)
✔ fixed normal exploration, pagination, SHA changes and repeated ranges are external observations, not a score (1.2676ms)
✔ read-only research is valid work; model summaries and unchanged timestamped listings do not become engineering evidence (0.5737ms)
✔ new permissions, network and broken tools are failure evidence; repeated invalid inputs remain candidates (0.378ms)
✔ waiting for original long checks, review and approvals stays waiting across repeated observations (0.4953ms)
✔ baseline facts, real repairs, current-source check results and requirement evidence have distinct growth (0.2021ms)
✔ bounded memory keeps unknown instead of claiming an unretained range is repeated; Task/Run instances cannot mix (12.5385ms)
✔ actual runtime records growing pages then a repeated page without changing model rounds, stop rules or original tool pairs (7091.4306ms)
✔ actual invalid tool calls repeat under the original step budget; observation never adds an early stop (6386.6521ms)
✔ actual command approval is observed as waiting and denial stays a permission failure (3953.3036ms)
✔ network errors remain model failures with shadow failure evidence, never a fabricated no-progress cause (4863.5938ms)
✔ observer receipt failure cannot change tool results, model budget or completion (3347.7248ms)
✔ a later receipt recovers after an observation write failure without misaligning retained source receipts (4782.6558ms)
✔ real applied files, counted checks and current requirement proofs grow after an existing-facts baseline (5013.8415ms)
✔ receipt retention stays aligned at the cap and a foreign step cannot enter the Run (3157.4456ms)
✔ an actual approved long check records its owned running job as waiting before the original counted result (5299.788ms)
ℹ tests 43
ℹ suites 0
ℹ pass 43
ℹ fail 0
ℹ cancelled 0
ℹ skipped 0
ℹ todo 0
ℹ duration_ms 75460.037