{ "round": "two-frontiers-cove-r1", "gate": "G-QUOTA", "verdict": "PASS", "detail": "the window is open", "started_utc": "2026-09-06T13:50:39Z", "finished_utc": "2026-09-06T13:50:41Z", "evidence": { "phase": "boundary", "arm": "cli-claude-fable-5-1", "wall_ms": 2007, "probe": { "collection_state": "COLLECTED", "rate_limit_receipt": { "status": "allowed_warning", "resetsAt": 1788742800, "rateLimitType": "seven_day_overage_included", "utilization": 0.9, "isUsingOverage": false, "surpassedThreshold": 0.75, "unifiedWindows": { "five_hour": { "utilization": 0.06, "resetsAt": 1788718200 }, "seven_day": { "utilization": 0.7, "resetsAt": 1788742800 }, "seven_day_overage_included": { "utilization": 0.9, "resetsAt": 1788742800 } } }, "cost": { "usd": 0.0028599999999999997, "state": "no-figure-held", "label": "list-rate estimate, not a receipt", "why": "this arm rides a consumer subscription (OAuth), so the CLI's own total_cost_usd is a LIST-RATE ESTIMATE of what the same tokens would have cost on the API, not a receipt for what was paid." }, "wall_ms": 2007 }, "a_session_limit_is_transport": "a session-limit reply is TRANSPORT and never content: it files NOT-COLLECTED — QUOTA, the arm resumes after the reset, and the gap prints." }, "arms": {}, "records": [ { "round": "two-frontiers-cove-r1", "leg": "probe", "arm_id": "cli-claude-fable-5-1", "model": "claude-fable-5-1", "transport_class": "agent-harness-cli", "dispatch_id": "g-quota", "attempt": 1, "collected": true, "collection_state": "COLLECTED", "collection_detail": null, "stamped_utc": "2026-09-06T13:50:41Z", "served_by_model": [ "claude-fable-5-1" ], "served_by_registered_model": true, "fallback": null, "prompt_sha256": "00101f21dea275accb9aa7ba015c66d1de62845820a4c43c71f7ffe614b15e1b", "system_sha256": "bc08b15e1d666817a85deb76df71182f5b17ee4ea314f8635a98fe14b17b4ad3", "response_sha256": "5fa7aac5375c5815787fba3f49559f9b45b14023147ce0652803387974144e5f", "response_text": "Ready", "response_chars": 5, "response_empty": false, "redaction": { "applied": false, "rule": "every email-shaped string and every bare account local part -> [REDACTED-EMAIL], at write time, in one function", "response_sha256_as_received": "5fa7aac5375c5815787fba3f49559f9b45b14023147ce0652803387974144e5f", "response_chars_as_received": 5, "note": "response_sha256 is over the PUBLISHED text, so a reader can verify it against the text in this row. response_sha256_as_received is over the bytes the arm returned; those bytes live only in the quarantined raw stream, so that hash is checkable on this box and nowhere else." }, "latency": { "class": "harness-wall", "ms": 2007, "harness_wall_ms": 2007, "duration_ms_reported": 1857, "duration_api_ms": 1847, "ttft_ms": 1823, "note": "harness-wall: a whole process spawn, not a socket round trip. Published as a labelled UPPER BOUND and never beside a serving figure (PLAN §8.2)." }, "usage": { "input_tokens": 261, "output_tokens": 5, "cache_creation_input_tokens": 0, "cache_read_input_tokens": 0, "thinking_tokens": 0, "thinking_tokens_reported": true, "service_tier": "standard", "note": "the CLI's own result.usage. thinking_tokens is None when no output_tokens_details block was reported and 0 when one was reported with a zero - G-EFFORT's receipt is '> 0', so the two must not collapse." }, "thinking": { "tokens": 0, "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855", "chars": 0, "blocks": 0, "text_stored": false, "why": "this arc records the LENGTH of a chain of thought and never its words; the raw stream is quarantined under results/raw-streams/ and excluded from the kit and from seal-manifest.json." }, "cost": { "usd": 0.0028599999999999997, "state": "no-figure-held", "label": "list-rate estimate, not a receipt", "why": "this arm rides a consumer subscription (OAuth), so the CLI's own total_cost_usd is a LIST-RATE ESTIMATE of what the same tokens would have cost on the API, not a receipt for what was paid." }, "law_record": { "seal": "agent-harness-cli", "argv_carries_the_prompt": false, "prompt_channel": "stdin (bytes)", "prompt_bytes": 10, "system_bytes": 19, "system_bytes_limit": 16384, "argv_element_wall_bytes": 131072, "system_channel": "argv (--system-prompt). Refused over 16,384 bytes: a single argv element dies at 131,072 on this box, measured, and execve raises before a process exists.", "effort": "high", "effort_channel": "argv only (--effort); no CLAUDE_* env var reaches the child", "sampler": "not settable on this transport (disclosed, never approximated)", "output_cap": "none settable", "g_id": { "asserted": [ "no `fallback` content block in the stream", "no `system/model_refusal_fallback` record in the stream", "every assistant `message.model` is the arm's registered model" ], "registered_model": "claude-fable-5-1", "served_by_model": [ "claude-fable-5-1" ], "unregistered_models": [], "fallback": null, "fallback_blocks": [], "model_refusal_fallbacks": [], "findings": [], "verdict": "PASS", "why": "identity is receipted per CALL, and a failed reading files THIS cell without stopping the arm (PREREG §3, A7): one swapped cell does not un-vouch its neighbours." }, "g_tools": { "asserted": [ "init.model", "init.tools == []", "init.mcp_servers == []", "init.slash_commands == []", "init.apiKeySource == 'none' (the OAuth/subscription value)", "no content block of ANY tool-shaped type, anywhere in the stream (census)", "no unknown content block type (fail-closed)", "result.usage.server_tool_use counters all zero", "result.permission_denials == []", "the scratch cwd is still empty after the call" ], "breaches": [], "unverifiable": [], "failures": [], "verdict": "PASS" }, "stream": { "records": 4, "unparsed_lines": 0, "block_type_census": { "text": 1 }, "known_block_types": [ "text", "thinking", "redacted_thinking", "fallback" ], "tool_shaped_block_types": [], "unknown_block_types": [], "string_content_records": 0, "unshaped_blocks": 0, "assistant_model_census": { "claude-fable-5-1": 1 }, "fallback_blocks": 0, "model_refusal_fallback_records": 0, "tool_use_blocks": 0, "server_tool_use": { "web_search_requests": 0, "web_fetch_requests": 0 }, "result_subtype": "success", "result_is_error": false, "stop_reason": "end_turn", "num_turns": 1, "api_key_source": "none", "rate_limit_status": "allowed_warning", "rate_limit_status_recognised": true, "rate_limit_statuses_known": [ "allowed", "allowed_warning", "rejected" ] } }, "artifacts": { "leak_markers_found": [], "reply_chars": 5, "reply_empty": false, "reply_sha256": "5fa7aac5375c5815787fba3f49559f9b45b14023147ce0652803387974144e5f" }, "cli_version": "2.1.263 (Claude Code)", "cwd_entries_after_call": [], "cwd_entries_after_call_asserted": true, "returncode": 0, "stderr_first_line": null, "init_receipt": { "type": "system", "subtype": "init", "model": "claude-fable-5-1", "tools": [], "mcp_servers": [], "slash_commands": [], "apiKeySource": "none", "permissionMode": "default", "claude_code_version": "2.1.263", "output_style": "default", "agents": [ "claude", "Explore", "general-purpose", "Plan" ], "skills": [], "plugins": [], "analytics_disabled": true, "product_feedback_disabled": true, "capabilities": [ "interrupt_receipt_v1", "interrupt_cancel_queued_v1", "msg_lifecycle_v1" ], "fast_mode_state": "off", "fast_mode_disabled_reason": "sdk_opt_in_required", "session_id": "68719346-7bb4-4adb-bbc9-f839c7bab900" }, "init_receipt_keys_published": [ "type", "subtype", "model", "tools", "mcp_servers", "slash_commands", "apiKeySource", "permissionMode", "claude_code_version", "output_style", "agents", "skills", "plugins", "analytics_disabled", "product_feedback_disabled", "capabilities", "fast_mode_state", "fast_mode_disabled_reason", "session_id", "cwd" ], "init_receipt_keys_dropped": [ "messaging_socket_path", "uuid" ], "init_receipt_note": "the init record, reduced to the keys this round publishes. The dropped keys are named, never withheld silently: memory_paths and messaging_socket_path carry the orchestrator's own session paths and are evidence about nothing the seal claims. The whole record survives in the quarantined raw stream.", "env_receipt": { "names_passed": [ "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC", "DISABLE_AUTOUPDATER", "DISABLE_BUG_COMMAND", "DISABLE_ERROR_REPORTING", "DISABLE_TELEMETRY", "HOME", "LANG", "PATH", "TERM", "USER" ], "flags": { "CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC": "1", "DISABLE_TELEMETRY": "1", "DISABLE_ERROR_REPORTING": "1", "DISABLE_AUTOUPDATER": "1", "DISABLE_BUG_COMMAND": "1" }, "allowlist": [ "PATH", "HOME", "USER", "LANG", "TERM" ], "extra_declared": [], "dropped_count": 73, "dropped_names": [ "AI_AGENT", "ATUIN_HISTORY_ID", "ATUIN_PREEXEC_BACKEND", "ATUIN_SESSION", "ATUIN_SHLVL", "ATUIN_TMUX_POPUP", "CLAUDECODE", "CLAUDE_CODE_CHILD_SESSION", "CLAUDE_CODE_ENTRYPOINT", "CLAUDE_CODE_EXECPATH", "CLAUDE_CODE_MESSAGING_SOCKET", "CLAUDE_CODE_MESSAGING_TOKEN", "CLAUDE_CODE_SESSION_ID", "CLAUDE_EFFORT", "CLAUDE_PID", "COLORTERM", "COREPACK_ENABLE_AUTO_PIN", "COVE_FENCE_PATH", "COVE_LOCAL_BASE_URL", "COVE_SEALED_DIR", "COVE_SEAL_SOURCE_ROOT", "DBUS_SESSION_BUS_ADDRESS", "DEBUGINFOD_URLS", "DESKTOP_SESSION", "DISPLAY", "GDMSESSION", "GHOSTTY_BIN_DIR", "GHOSTTY_RESOURCES_DIR", "GHOSTTY_SHELL_FEATURES", "GIT_EDITOR", "GNOME_DESKTOP_SESSION_ID", "GNOME_KEYRING_CONTROL", "GNOME_SETUP_DISPLAY", "GPG_AGENT_INFO", "GTK_MODULES", "HISTCONTROL", "IM_CONFIG_ENTRY", "LESSCLOSE", "LESSOPEN", "LOGNAME", "LS_COLORS", "MANAGERPID", "MANAGERPIDFDID", "MCP_TIMEOUT", "MEMORY_PRESSURE_WATCH", "MEMORY_PRESSURE_WRITE", "NoDefaultCurrentDirectoryInExePath", "PWD", "PYTHONDONTWRITEBYTECODE", "QT_ACCESSIBILITY", "QT_IM_MODULE", "QT_IM_MODULES", "SHELL", "SHLVL", "SSH_AUTH_SOCK", "SYSTEMD_EXEC_PID", "TERMINFO", "TERM_PROGRAM", "TERM_PROGRAM_VERSION", "USERNAME", "WAYLAND_DISPLAY", "XAUTHORITY", "XDG_CONFIG_DIRS", "XDG_CURRENT_DESKTOP", "XDG_DATA_DIRS", "XDG_MENU_PREFIX", "XDG_RUNTIME_DIR", "XDG_SESSION_CLASS", "XDG_SESSION_DESKTOP", "XDG_SESSION_EXTRA_DEVICE_ACCESS", "XDG_SESSION_TYPE", "XMODIFIERS", "_" ], "note": "an ALLOWLIST, derived from the env this driver actually passed. G-EGRESS receipts the ENV alongside the host list: an interactive-grade CLI does analytics, error reporting, plugin-stats fetches and an autoupdate check unless the five flags are set -- and inherits an IPC socket, a session id, an effort override and any exported API key unless the environment is an allowlist. NAMES are recorded, never values." }, "rate_limit_receipt": { "status": "allowed_warning", "resetsAt": 1788742800, "rateLimitType": "seven_day_overage_included", "utilization": 0.9, "isUsingOverage": false, "surpassedThreshold": 0.75, "unifiedWindows": { "five_hour": { "utilization": 0.06, "resetsAt": 1788718200 }, "seven_day": { "utilization": 0.7, "resetsAt": 1788742800 }, "seven_day_overage_included": { "utilization": 0.9, "resetsAt": 1788742800 } } }, "truthful_path_fields": [ "cwd", "cli_executable", "raw_stream_path", "attempt_file" ], "truthful_path_fields_round": [ "bank", "argv", "cwd", "cli_executable", "raw_stream_path", "attempt_file", "sent_file", "base_url", "planted_at", "path", "projects_dir", "prompt_path", "receipts_dir", "receipt_path", "bank", "queries", "fixture", "roster", "receipt", "legC_filler", "_filler", "rows_path", "previous_rows_kept_at" ], "run": "scored" } ], "withheld": { "rule": "a receipt publishes in the kit, so these fields are removed here rather than at kit time — and named, because a field quietly absent from a receipt is indistinguishable from a gate that never read it." }, "note": "written by harness/cove/gates_cove.py. Every generation call rides the same send() the scored run uses; a gate that hand-rolled its own call would receipt a field the run never sent." }