@songtonyli/dsh-cli 0.1.62 → 0.1.63
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +1 -1
- package/bin/dsh.mjs +1 -1
- package/node_modules/@deepseek-ai/cordis/package.json +4 -4
- package/node_modules/@deepseek-ai/cordis-plugin-group/package.json +3 -3
- package/node_modules/@deepseek-ai/cordis-plugin-include/package.json +4 -4
- package/node_modules/@deepseek-ai/cordis-plugin-loader/package.json +4 -4
- package/node_modules/@deepseek-ai/cordis-plugin-timer/package.json +3 -3
- package/node_modules/@deepseek-ai/cosmokit/package.json +1 -1
- package/node_modules/@deepseek-ai/dsh/package.json +117 -117
- package/node_modules/@deepseek-ai/dsh-acp/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-acp-app/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-agent/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-agent-default-model/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-agent-instructions/package.json +26 -26
- package/node_modules/@deepseek-ai/dsh-agent-loop/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-agent-preset/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-agent-preset-registry/package.json +34 -34
- package/node_modules/@deepseek-ai/dsh-agent-tool-presentation/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-anonymous-user-id/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-api-account-controller/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-api-authorization-controller/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-api-gateway/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-api-job-controller/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-api-remotes/package.json +42 -42
- package/node_modules/@deepseek-ai/dsh-api-session-controller/package.json +72 -72
- package/node_modules/@deepseek-ai/dsh-api-settings-controller/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-api-terminal-controller/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-api-workspace-controller/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-api-workspace-files/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-app-boot/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-atomic-write/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-attachment/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-attachment-local/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-authorization/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-base/package.json +96 -96
- package/node_modules/@deepseek-ai/dsh-bash-local/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-bash-sandbox/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-brand/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-chunked-list/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-client-connection/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-client-file-upload/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-client-hmr/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-client-locale/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-client-modules/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-client-product-analytics/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-client-resources/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-client-shortcuts/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-store/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-client-ui-agent-preset/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-client-ui-approval/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-client-ui-attachment/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-brand-official/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-client-ui-chat/package.json +40 -40
- package/node_modules/@deepseek-ai/dsh-client-ui-commands/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-client-ui-conversation/package.json +36 -36
- package/node_modules/@deepseek-ai/dsh-client-ui-cordis/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-client-ui-deliverables/package.json +30 -30
- package/node_modules/@deepseek-ai/dsh-client-ui-directory-picker-browse/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-client-ui-directory-picker-native/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-client-ui-goal/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-client-ui-input-trigger/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-client-ui-jobs/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-client-ui-layout/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-message-feedback/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-client-ui-model-selection/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-client-ui-open-in-app/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-client-ui-permission-presets/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-client-ui-plan/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-client-ui-plugin-manager/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-client-ui-primitives/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-client-ui-reference/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-client-ui-renderer/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-client-ui-schedule/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-client-ui-session/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-client-ui-settings/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-account/lib/client.js +2 -2
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-account/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-agent-loop/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-general/lib/client.js +1 -1
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-general/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-models/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-plugin-inventory/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-plugins/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-session-log/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-shell/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-signin/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-subagent/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-settings-web-search/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-client-ui-shortcuts/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar/lib/client.js +1 -1
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar-browser/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar-documentpreview/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar-files/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar-right/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-client-ui-sidebar-terminal/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-client-ui-skill/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-client-ui-slots/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-client-ui-subagent/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-client-ui-theme/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-client-ui-tool/package.json +27 -27
- package/node_modules/@deepseek-ai/dsh-client-ui-trajectory/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-client-ui-user-questions/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-client-ui-workflow-run/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-client-ui-workspace/package.json +26 -26
- package/node_modules/@deepseek-ai/dsh-cmdline/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-command-compact/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-command-feedback/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-command-goal/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-commands/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-compaction/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-compaction-basic/package.json +26 -26
- package/node_modules/@deepseek-ai/dsh-compaction-image-offload/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-compaction-tool-result-pruner/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-config-editor/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-cordis-client-runner/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-cordis-host-runner/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-credentials/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-credentials-local/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-deepseek-account/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-deepseek-account-platform/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-deepseek-llm-api-extensions/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-deque/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-experimental-agent-team/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-experimental-agent-team-profile/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-experimental-api-speech-to-text/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-experimental-auto-review/package.json +41 -41
- package/node_modules/@deepseek-ai/dsh-experimental-client-ui-agent-team/package.json +24 -24
- package/node_modules/@deepseek-ai/dsh-experimental-client-ui-voice-input/package.json +33 -33
- package/node_modules/@deepseek-ai/dsh-experimental-inspector/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-experimental-inspector-profile/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-experimental-session-inspector/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-experimental-speech-to-text/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-experimental-speech-to-text-sensevoice/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-experimental-tool-agent-team/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-experimental-voice-input-bundle/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-file-reference/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-file-reference-local/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-fs/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-fs-local/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-fs-observation-policy/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-fs-sandbox/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-goal/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-goal-round-driver/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-headless/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-hmr/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-home-paths/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-hook-protocol/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-hooks-claude-code/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-hooks-codex/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-host-directory-picker/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-host-directory-picker-auto/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-host-directory-picker-browse/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-host-directory-picker-native/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-host-frontend-static/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-host-open-in-app/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-host-plugin-inventory/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-host-product-telemetry-otel/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-host-webserver/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-http-proxy/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-jobs/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-jobs-local/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-launch-environment/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-lazy-require/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-llm/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-llm-cursor/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-llm-deepseek/package.json +44 -44
- package/node_modules/@deepseek-ai/dsh-llm-deepseek-account/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-llm-deepseek-api-key/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-llm-pi-ai/package.json +26 -26
- package/node_modules/@deepseek-ai/dsh-llm-retry/package.json +24 -24
- package/node_modules/@deepseek-ai/dsh-mcp-client/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-mcp-resources/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-message-feedback/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-native-command/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-no-progress-reminder/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-office-to-pdf/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-otel/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-output-retention/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-package-manifest/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-permission-presets/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-persona/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-plan-mode/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-plugin-manager/package.json +33 -33
- package/node_modules/@deepseek-ai/dsh-plugin-package-inventory-deepseek/package.json +19 -19
- package/node_modules/@deepseek-ai/dsh-ptc-runtime/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-ptc-runtime-node/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-pwsh-local/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-pwsh-sandbox/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-repeat-tool-reminder/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-sandbox/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-sandbox-apple-container/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-sandbox-local/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-sandbox-policy/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-sandbox-windows-acl/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-schedule/package.json +34 -34
- package/node_modules/@deepseek-ai/dsh-scope/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-sdk-app/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-sdk-jsonrpc-server/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-sdk-minimal/package.json +29 -29
- package/node_modules/@deepseek-ai/dsh-sdk-protocol/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-session/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-session-checkpoint-policy/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-session-format/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-session-format-catalog/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-session-format-v0-to-v1/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-session-format-v1-to-v2/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-session-format-v2-to-v3/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-session-format-v3-to-v4/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-session-log-deepseek/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-session-log-export/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-session-persistence/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-session-persistence-jsonl/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-session-projection/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-session-projection-cache/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-session-query/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-session-query-sqlite/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-session-reference/package.json +39 -39
- package/node_modules/@deepseek-ai/dsh-session-stats/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-session-telemetry/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-session-telemetry-otel/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-session-title/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-session-title-first-prompt-llm/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-session-title-llm/package.json +13 -13
- package/node_modules/@deepseek-ai/dsh-session-turn-outline/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-settings/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-shell/package.json +7 -7
- package/node_modules/@deepseek-ai/dsh-shell-env/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-skill/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-skill-badge/package.json +5 -5
- package/node_modules/@deepseek-ai/dsh-skill-filesystem/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-skill-office/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-spill/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-spill-local/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-spill-policy/package.json +33 -33
- package/node_modules/@deepseek-ai/dsh-storage/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-storage-domain/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-storage-json/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-subagent/package.json +45 -45
- package/node_modules/@deepseek-ai/dsh-subagent-acp/package.json +24 -24
- package/node_modules/@deepseek-ai/dsh-subagent-fork-in-process/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-subagent-in-process-driver/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-subagent-spawn-in-process/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-subprocess/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-subprocess-local/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-system-prompt/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-terminal/package.json +8 -8
- package/node_modules/@deepseek-ai/dsh-terminal-bash/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-time-context/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-timeout/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-tmux-context/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-token-meter/package.json +17 -17
- package/node_modules/@deepseek-ai/dsh-tool-ask-user/package.json +12 -12
- package/node_modules/@deepseek-ai/dsh-tool-bash/package.json +32 -32
- package/node_modules/@deepseek-ai/dsh-tool-bash-persistent/package.json +24 -24
- package/node_modules/@deepseek-ai/dsh-tool-call-timeout-policy/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-tool-cordis/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-tool-fs/package.json +32 -32
- package/node_modules/@deepseek-ai/dsh-tool-fs-search/package.json +28 -28
- package/node_modules/@deepseek-ai/dsh-tool-goal/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-tool-jobs/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-tool-present/package.json +20 -20
- package/node_modules/@deepseek-ai/dsh-tool-pwsh/package.json +31 -31
- package/node_modules/@deepseek-ai/dsh-tool-pwsh-persistent/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-tool-ralph/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-tool-schedule/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-tool-skill/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-tool-str-replace-editor/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-tool-subagent/package.json +28 -28
- package/node_modules/@deepseek-ai/dsh-tool-subagent-control/package.json +21 -21
- package/node_modules/@deepseek-ai/dsh-tool-todo/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-tool-web/package.json +22 -22
- package/node_modules/@deepseek-ai/dsh-tool-workflow/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-tool-workspace-dependencies/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-tools/package.json +25 -25
- package/node_modules/@deepseek-ai/dsh-tui-app/README.i18n.yaml +6 -6
- package/node_modules/@deepseek-ai/dsh-tui-app/README.md +6 -5
- package/node_modules/@deepseek-ai/dsh-tui-app/README.zh.md +6 -5
- package/node_modules/@deepseek-ai/dsh-tui-app/lib/index.js +396 -111
- package/node_modules/@deepseek-ai/dsh-tui-app/package.json +81 -81
- package/node_modules/@deepseek-ai/dsh-typert-loader/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-typert-protocol/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-typert-registry/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-user-approval/package.json +16 -16
- package/node_modules/@deepseek-ai/dsh-user-questions/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-util-code-language/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-util-crypto/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-util-time/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-util-values/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-util-workspace-path/package.json +3 -3
- package/node_modules/@deepseek-ai/dsh-web/package.json +6 -6
- package/node_modules/@deepseek-ai/dsh-web-app/package.json +146 -146
- package/node_modules/@deepseek-ai/dsh-web-fetch-http/package.json +10 -10
- package/node_modules/@deepseek-ai/dsh-web-frontend/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-web-search-deepseek/package.json +18 -18
- package/node_modules/@deepseek-ai/dsh-web-search-exa/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-web-search-perplexity/package.json +9 -9
- package/node_modules/@deepseek-ai/dsh-webhook/package.json +23 -23
- package/node_modules/@deepseek-ai/dsh-webhook-github/package.json +15 -15
- package/node_modules/@deepseek-ai/dsh-win32-process/package.json +4 -4
- package/node_modules/@deepseek-ai/dsh-workflow/package.json +11 -11
- package/node_modules/@deepseek-ai/dsh-workflow-ptc/package.json +36 -36
- package/node_modules/@deepseek-ai/dsh-workspace/package.json +14 -14
- package/node_modules/@deepseek-ai/dsh-workspace-changes/package.json +16 -16
- package/node_modules/@deepseek-ai/schemastery/package.json +2 -2
- package/package.json +291 -291
|
@@ -40,7 +40,7 @@ On quit the app prints `dsh: session <id> saved; resume with: dsh --profile tui
|
|
|
40
40
|
|
|
41
41
|
### The screen
|
|
42
42
|
|
|
43
|
-
Every frame is drawn one column in from each edge of the terminal, so no line of the conversation or of the docked chrome touches either side. The header names the session by its title once one is generated or set, with the id beside it. The transcript grows in the terminal's own scrollback: your prompts start with `❯` on a tinted full-width band while colour is on (attachments listed under them), and a prompt steered or injected into a turn that already had a model step draws dim on a darker band so it recedes behind the turn it joined, assistant reasoning is dim italic under a `✻ Thinking` header above the Markdown reply, and each tool call is a card with a status glyph — a yellow
|
|
43
|
+
Every frame is drawn one column in from each edge of the terminal, so no line of the conversation or of the docked chrome touches either side. The header names the session by its title once one is generated or set, with the id beside it. The transcript grows in the terminal's own scrollback: your prompts start with `❯` on a tinted full-width band while colour is on (attachments listed under them), and a prompt steered or injected into a turn that already had a model step draws dim on a darker band so it recedes behind the turn it joined, assistant reasoning is dim italic under a `✻ Thinking` header above the Markdown reply, and each tool call is a card with a status glyph — a yellow spinner of that tool family's flowers, blooms, and stars while it runs, a green `◆` once done, a red `◆` on failure — the bold tool name, the presenter headline in violet — without the tool name a presenter repeated, so `Read src/app.ts` draws as `read src/app.ts`, and for a shell card the call's own summary, never the command, which the card draws once as its `$` row — a loading row while it runs, and a body folded to `toolPreviewLines` rows. The card appears when the model starts the call, its header first and its rows unrolling beneath it over about `toolRevealFrames` frames; the result rows unroll and float out the same way when the tool answers. An `edit` or `write` diff card fills each added row edge to edge with a green background and each removed row with a red one, the removed rows above the added ones, and paints the `+` and `-` in their own lighter colour so each sign reads against the fill rather than in it; a wrapped row carries the fill on without a second sign; once the tool applied the change, the applied hunks replace the diff the call carried and lead every row with its file line number, the new file's for added and unchanged rows and the old file's for removed rows. A `subagent` or `subagent_*` card draws folded as one row behind the same status glyph, spinning while it runs and a dim `◆` for a child handed to the background: while it runs, the tool name, the task description, and the requested model and `background` when the call names them; once settled, the description dimmed with `[done]`, `[started]` for a child handed to the background, or `[failed]` at the right edge. `Space` on the marked row, or `Ctrl+O`, unfolds the full card. A nonempty system prompt and each injected context — instructions, catalogs, snapshots, notices, relays, and recalls — draw as a dim `⬡` title over the first `contextPreviewLines` rows of their model-facing text, with each snapshot contribution named above its own rows; an empty system prompt is omitted. A compaction, from `/compact` or automatic, draws as one context block whose accent `⬡` title names its origin and the estimated tokens it replaced and its summary costs, as in `context compacted by /compact · ~6.4k → ~0.9k tokens` or `context compacted automatically in turn 12 · …`; under the title, `compressed` names the item count and turns the summary replaced, each prompt among them (five at most, one row each), and a count of the replies, tool results by tool, context blocks, and earlier summaries, and `preserved` names the system prompt and what each later turn still holds verbatim; both stay drawn while the summary under them folds to `contextPreviewLines` rows like any context block. The checkpoint message itself is not drawn again. A `/compact` attempt the log closes with an error prints `/compact failed · <error>` when the session is redrawn, since the live command reports its own failure; a failed automatic attempt changes nothing and draws nothing, as in the browser. A tool card and a context block that carry more rows than they draw end on one marker: `… <n> more rows · Space expands` while the conversation's focus holds that block, and `… <n> more rows · Ctrl+O expands` otherwise. Fenced code in a reply, the file rows of a `read` card, the changed rows of an `edit` or `write` diff card — with `+` additions marked green, `-` removals red, and source text behind those signs drawn in the language the file's extension names — and the shell command of a `!` / `!!` draft, a `$ command` user-shell row (its `$` in the accent colour), and a terminal tool card, use syntax colours from a theme the terminal's own background picks, light or dark; the grammar of a language loads on the first block that asks for it, so that block draws plain and is repainted in colour a moment later, and a language with no grammar here, a terminal that reports neither 24-bit nor 256 colours, and `codeHighlight: false` each draw every block plain. Headings and list markers use warm orange, links and inline code use violet, and quotes stay dim. Streamed reasoning, reply text, and tool arguments are queued as they arrive and drawn a share per frame, about 60 frames per second by default, so a network burst spreads over `streamPaceFrames` frames instead of landing at once, and the queue never holds the model call back. Streamed reply text fades in and every word keeps its own clock: a word enters near the terminal's background color and brightens to the color it settles in over `streamFadeSteps` ticks of `streamFadeStepMs` each, so a fast stream leaves a longer trail of brightening words rather than a darker one. Streamed reasoning and a tool card float out over the same duration: they appear at a lifted color and recede to the dim italic or palette colors they settle in; a card redrawn from persisted history carries no fade, and text that has settled is never dimmed again.
|
|
44
44
|
|
|
45
45
|
Below the transcript sit a spinner while the agent works — `thinking`, `writing`, `calling <tool>`, a retry, or `compacting`, each beside its own braille animation: circling dots, filling lines, an orbiting dot, a sliding bar, or a settling cell — with a bright highlight sweeping across the word, plus live `↑` send and `↓` receive tokens for the current model call, as compact `k` / `M` / `B` counts, growing as the stream arrives and settling on the provider usage chunk — any open prompt, a boxed follow-ups list while messages wait, an activity board of the open turn's todos and latest descendant line while either has a row — completed todo content scratched out and status glyphs colored green, yellow, or cyan, the in-progress one a spinner, at most four todo rows then `+<n> more`, not a focus region, gone on `turn/end` and bind — the editor, the subagent panel, and the footer: a status bar that, unfocused, is one line built from both ends inward. The editor's caret is the terminal's own blinking bar: the app asks for that shape at start, gives your default back on quit, and draws no caret at all while the status bar or the panel holds the keyboard. The model and entry keys are colored; live turn, context, and todo facts use yellow, violet, and green. The model is anchored at the left edge and is never dropped — past 20 columns it falls back to the bare model name, and only a label still too wide for that is ellipsized — and the two keys that leave the input are anchored at the right edge as `Shift+↑ read · Shift+↓ status`, or `Shift+↑ select · Shift+↓ status` while follow-ups wait, narrowing to `Shift+↑↓ nav` and dropped below 40 columns. The key facts fill whatever middle is left, in bar order: effort (`effort default` when the selection leaves reasoning to the model), the running turn's elapsed time, the context window percentage, the todo counts, and the workspace path (shortened with `~` and `…/` when long). Every segment that middle cannot hold folds into a trailing `+N`, together with the segments the unfocused line never draws — the permission preset, token usage (session totals plus the in-flight call), the goal and plan-mode markers from the projection seam, and the count of pending attachments. Model-request retries appear as notices, the same facts the browser's markers carry.
|
|
46
46
|
|
|
@@ -48,11 +48,11 @@ While the keyboard walks the conversation a framed inspector is docked directly
|
|
|
48
48
|
|
|
49
49
|
The focused block is also marked where it stands, with a two-column gutter that is dim beside the block's other lines and accented beside the focused section's own — or beside the fold marker, when the fold left that section out of the drawn rows entirely — and its content wraps two columns narrower while the mark is drawn. That mark reaches only the lines pi-tui still repaints differentially — the last `rows` lines of the last frame it wrote, a window a taller frame raises and a shorter one never lowers — because changing anything above them would clear the terminal's scrollback. A block that has scrolled past that point keeps no gutter, the inspector heading appends ` · off screen`, and the frame's own rules draw dim instead of accent, so the frame itself reports that the mark cannot be drawn where the block stands. A fade whose rows reach that edge settles them to their final colors in the frame that writes them, so nothing stays dim in the scrollback.
|
|
50
50
|
|
|
51
|
-
`Ctrl+G` from any region, `Enter` on a conversation section, and `/turns` read the conversation full screen, one turn at a time. The reader takes the terminal rather than drawing over the conversation: it runs on the terminal's alternate screen, the screen a pager runs on, and the conversation's own screen is held off the terminal for as long as it is up. So it costs the conversation no line, leaves the repaint window where it was, fills the screen on every terminal, and cannot leave a row of itself in your scrollback; closing it restores the conversation exactly as you left it, with whatever landed meanwhile drawn underneath. When the turn panel scrolls below its first row, a compact tinted bar replaces the top rule with the current user prompt, collapsed to one line and ellipsized to the terminal width; scrolling back to the top restores the ` ● READER ` chip and `turn <n> of <total>`. Under that fixed header, the turn being read takes most of the screen and a narrow turn list sits to its left, every body row closed by a right border. The list has one row per turn — `❯` on the turn being read, the turn number, the first line of its prompt, and right-aligned markers, `⬡<n>` injected context blocks, `✻` any reasoning, `¶` any reply, `◆<n>` tool calls — with the sections of the turn being read listed under it, `▸` on the one the reader holds, so any turn's own context is one step away. The turn panel scrolls that turn end to end in the conversation's own glyphs, a blank row between each two sections: the prompt opens it behind `❯` on a shaded band, then `✻ Thinking` in dim italic, `¶ Reply` rendered as the conversation renders Markdown — headings, emphasis, inline code, and fenced code in syntax colour — except that unordered bullets are drawn as `•`, a call's `◆` — green once its result landed, yellow while it runs — with the tool name in bold and its headline in the link blue, `⎿ Result` indented under its call with a diff's `+ ` rows green and `- ` rows red on shaded bands and a `read` card's file rows in syntax colour, and `⬡` with the label of an injected context. The section the top row belongs to is accented and carries the conversation walk's `┃` gutter on every row, which reads with no color at all. The bands are mixed from the background color the terminal reports, so a terminal that reports none, or draws no color, draws the panel without them. Rows are never folded there: the reader is where the full model-facing text of a system prompt, an injected snapshot, a long reply, or a tool result is read. `Right` opens the selected turn and `Left` goes back to the list; in the list `Up` / `Down` step one section at a time and cross into the turn either side while `PageUp` / `PageDown` step a whole turn, in the turn they scroll one row at a time, and the section the top row belongs to is where the conversation resumes when the reader closes. `/` opens a query line in place of the legend and narrows the list to the turns whose prompt or text matches, with the readout reading `<kept>/<total> turns`; otherwise the legend rule reports `turn 3/12 · row 14/212` at its right, which no width drops. Below `readerMinColumns` the body draws one panel at a time — the list, or the turn `Right` opens from it — and below 24 columns or 8 rows it draws one dim line, `terminal too small for the reader (needs 24×8)`, and answers only `Esc`, `Ctrl+G`, and `Ctrl+C`. It re-reads the conversation on every render, so a streaming reply grows inside it, a landing tool result appears, and a new turn joins the list; a rewrap re-anchors on the section being read rather than on a row number; an approval or a question steps it aside for that prompt's lifetime; and a session switch takes it down with the notice `the transcript changed · reader closed`.
|
|
51
|
+
`Ctrl+G` from any region, `Enter` on a conversation section, and `/turns` read the conversation full screen, one turn at a time. The reader takes the terminal rather than drawing over the conversation: it runs on the terminal's alternate screen, the screen a pager runs on, and the conversation's own screen is held off the terminal for as long as it is up. So it costs the conversation no line, leaves the repaint window where it was, fills the screen on every terminal, and cannot leave a row of itself in your scrollback; closing it restores the conversation exactly as you left it, with whatever landed meanwhile drawn underneath. When the turn panel scrolls below its first row, a compact tinted bar replaces the top rule with the current user prompt, collapsed to one line and ellipsized to the terminal width; scrolling back to the top restores the ` ● READER ` chip and `turn <n> of <total>`. Under that fixed header, the turn being read takes most of the screen and a narrow turn list sits to its left, every body row closed by a right border. The list has one row per turn — `❯` on the turn being read, the turn number, the first line of its prompt, and right-aligned markers, `⬡<n>` injected context blocks, `✻` any reasoning, `¶` any reply, `◆<n>` tool calls — with the sections of the turn being read listed under it, `▸` on the one the reader holds, so any turn's own context is one step away. The turn panel scrolls that turn end to end in the conversation's own glyphs, a blank row between each two sections: the prompt opens it behind `❯` on a shaded band, then `✻ Thinking` in dim italic, `¶ Reply` rendered as the conversation renders Markdown — headings, emphasis, inline code, and fenced code in syntax colour — except that unordered bullets are drawn as `•`, a call's `◆` — green once its result landed, yellow while it runs — with the tool name in bold and its headline in the link blue, `⎿ Result` indented under its call with a diff's `+ ` rows green and `- ` rows red on shaded bands and a `read` card's file rows in syntax colour, and `⬡` with the label of an injected context. The section the top row belongs to is accented and carries the conversation walk's `┃` gutter on every row, which reads with no color at all. The bands are mixed from the background color the terminal reports, so a terminal that reports none, or draws no color, draws the panel without them. Rows are never folded there: the reader is where the full model-facing text of a system prompt, an injected snapshot, a long reply, or a tool result is read. `Right` opens the selected turn and `Left` goes back to the list; in the list `Up` / `Down` step one section at a time and cross into the turn either side while `PageUp` / `PageDown` step a whole turn, in the turn they scroll one row at a time, and the section the top row belongs to is where the conversation resumes when the reader closes. `/` opens a query line in place of the legend and narrows the list to the turns whose prompt or text matches, with the readout reading `<kept>/<total> turns`; otherwise the legend rule reports `turn 3/12 · row 14/212` at its right, which no width drops. Below `readerMinColumns` the body draws one panel at a time — the list, or the turn `Right` opens from it — and below 24 columns or 8 rows it draws one dim line, `terminal too small for the reader (needs 24×8)`, and answers only `Esc`, `Ctrl+G`, and `Ctrl+C`. It re-reads the conversation on every render, draws only sections in view, and reuses cached rows when their content, width, and styling match, so long turns stay responsive while the session runs; a streaming reply grows inside it, a landing tool result appears, and a new turn joins the list; a rewrap re-anchors on the section being read rather than on a row number; an approval or a question steps it aside for that prompt's lifetime; and a session switch takes it down with the notice `the transcript changed · reader closed`.
|
|
52
52
|
|
|
53
53
|
Transient key feedback floats as a dim framed line in the top-right corner of the viewport: `press Esc again to stop turn <n>`, `press Ctrl+C again to quit`, and `nothing in the transcript to read yet`. It holds at full strength for `toastMs`, fades out, costs the conversation no row of its own, and takes no keyboard; where the renderer can no longer repaint the top of the viewport it is printed into the conversation instead, which keeps it. Facts worth keeping are notices in the conversation: `stopping the turn…`, `wait for the session switch to finish`, `above the repaint window · opened in the reader`, and every command result.
|
|
54
54
|
|
|
55
|
-
Motion reaches past streamed text: every lift and float rides the same fade tick, and every working animation one spinner tick. The keyboard landing on the conversation, the subagent panel, or the status bar lifts that surface's frame, its chip, and the mark it holds for twelve ticks of `streamFadeStepMs`; a step of the conversation walk lifts the newly focused gutter for six; and a step along the bar lifts the selected label for eight. A new or status-changed activity-board todo row, and a replaced descendant line, float out on the fade tick the same way a tool card does; replay and `reducedMotion` draw the settled board. Every working animation has eight frames, takes the one it draws from the clock, and steps on the spinner tick of `spinnerMs`, 100 ms by default, so everything that spins advances at the same instant and speed, a label change keeps the beat, and the tick is armed only while one of them is drawn. The working spinner alone draws braille, its animation following its status, so it never repeats a card's glyph.
|
|
55
|
+
Motion reaches past streamed text: every lift and float rides the same fade tick, and every working animation one spinner tick. The keyboard landing on the conversation, the subagent panel, or the status bar lifts that surface's frame, its chip, and the mark it holds for twelve ticks of `streamFadeStepMs`; a step of the conversation walk lifts the newly focused gutter for six; and a step along the bar lifts the selected label for eight. A new or status-changed activity-board todo row, and a replaced descendant line, float out on the fade tick the same way a tool card does; replay and `reducedMotion` draw the settled board. Every working animation has eight frames, takes the one it draws from the clock, and steps on the spinner tick of `spinnerMs`, 100 ms by default, so everything that spins advances at the same instant and speed, a label change keeps the beat, and the tick is armed only while one of them is drawn. The working spinner alone draws braille, its animation following its status, so it never repeats a card's glyph. A running tool card and a folded subagent row draw the animation of that tool's family, and an in-progress todo row draws the planning family's. Each of those animations mixes flowers, blooms, and stars, no two families share a glyph, and every frame of one family is different, so two indicators of one family at different phases never draw the same glyph. An indicator that starts while others of its family spin takes a phase at which it never draws their glyph at the same instant, for up to eight of that family spinning together; beyond eight, it takes the phase the fewest indicators of that family hold. A card whose header has scrolled above the part of the screen the renderer can repaint draws the static `◆` from then on, and `reducedMotion` draws every one of them still — the card's `◆`, the todo's `▸`, and the working spinner's label without a glyph — with no spinner tick. The reader arrives complete on the screen it owns, and what it newly shows floats out on the fade tick the way a tool card does: opening it and showing another turn float out the whole turn, and stepping the list to another section of the same turn floats out that section, over `streamFadeSteps` ticks; scrolling, live growth, and a rewrap float nothing out. No lift changes how many lines anything draws, and `reducedMotion` — like `NO_COLOR`, a disabled palette, and `TERM=dumb` — turns every lift and float off and arms no repaint for them.
|
|
56
56
|
|
|
57
57
|
The subagent panel is drawn while a subagent session under the bound one is resident, or the listing carries a candidate it could not read. Unfocused it is one summary line with the listed count and the first child's key label; focused, its heading counts what it lists and ends on the keys it answers, `↑↓ children · Enter opens · Tab regions · Esc input`, narrowed to `↑↓ children · Esc input` and then to `Esc input` on a terminal that cannot hold them, and each row gives the child's depth indent, its label or id, its mode (`one-shot` or `continuable`), `resident`, whether its agent is `running` or `idle`, its elapsed time — the open turn's, else the total its settled turns took — and its token usage, as far as this process's own view of that child and the composed projections carry them. Six rows are drawn at most, with `+<n> more · /subagents lists them all` under them; a candidate the listing could not interpret draws as `unreadable: <reason>` and opens nothing, and a failed listing keeps the rows the last good one produced with `listing failed: <reason>` beneath. The panel disappears with its last row. `Enter` or `Right` on a readable row, like `Enter` on a readable `/subagents` row, opens that child as a subagent view: the terminal draws the child session in place of the bound one — its transcript, live stream, conversation walk, inspector, reader, and its own subagent panel — while the session it was entered from stays open. A resident child is read live, and any other is resumed. The entry's detail rows print as notices under `subagent <label> · Ctrl+P or /parent returns to session <id>`, the header ends on `◆ subagent view ›` with one `›` per level entered and `· Ctrl+P returns`, a shaded line directly above the input reads `◆ subagent view · main › <label> · Ctrl+P back to <parent>` for as long as the view is open, whatever the transcript has scrolled to, and `Ctrl+P` or `/parent` goes back one level, redrawing the parent from its log as it stands then. `/new`, `/clear`, `/resume`, `/sessions`, and `/fork` refuse while any session in that chain runs a turn, and otherwise release every view with the session they leave; quitting releases every view and saves the root session.
|
|
58
58
|
|
|
@@ -306,7 +306,7 @@ The patch rides over `dsh-base`: it sets the coding persona prefix and cwd suffi
|
|
|
306
306
|
| [`src/fade.ts`](src/fade.ts) | The streamed-text fade: the wall-clock tail tracker, the block-fade clock and registry, the fade-in ramp, the float-out mix, and the recolor of rendered lines |
|
|
307
307
|
| [`src/pace.ts`](src/pace.ts) | The stream pacer: the ordered queue between the assistant stream and the transcript, drained a backlog share per frame; and the row reveal that unrolls a tool card the same way |
|
|
308
308
|
| [`src/motion.ts`](src/motion.ts) | The chrome motion clock and the three-level lift its call sites draw with: the landing, the section step, and the bar walk |
|
|
309
|
-
| [`src/spinner.ts`](src/spinner.ts) | The working animations: the activity line's braille status animations, the star animation of running cards, subagent rows, and in-progress todos, the frame each draws at one instant, the phase that keeps
|
|
309
|
+
| [`src/spinner.ts`](src/spinner.ts) | The working animations: the activity line's braille status animations, the flower, bloom, and star animation of each tool family for running cards, subagent rows, and in-progress todos, which tool draws which, the frame each draws at one instant, the phase that keeps one family's indicators apart, and the shimmer the activity label carries |
|
|
310
310
|
| [`src/prompts.ts`](src/prompts.ts) | Approval, question, picker, and read-only detail prompts plus the modal queue |
|
|
311
311
|
| [`src/toast.ts`](src/toast.ts) | The transient key-feedback line: its overlay, its clock, and the lines it carries |
|
|
312
312
|
| [`src/pinned-prompt.ts`](src/pinned-prompt.ts) | The prompt row the reader pins over a scrolled turn: the glyph, the prompt collapsed to one line, and the band |
|
|
@@ -338,7 +338,7 @@ The patch rides over `dsh-base`: it sets the coding persona prefix and cwd suffi
|
|
|
338
338
|
| [`tests/editor.spec.ts`](tests/editor.spec.ts) | The terminal caret, `Shift+Left` / `Shift+Right` word navigation, and `!` / `!!` draft colour against pi-tui's editor behavior |
|
|
339
339
|
| [`tests/frame.spec.ts`](tests/frame.spec.ts) | The rules, the chip, the body rows, and the legend a width holds |
|
|
340
340
|
| [`tests/motion.spec.ts`](tests/motion.spec.ts) | The motion clock's levels, its repaint demand, and the lift each level draws |
|
|
341
|
-
| [`tests/spinner.spec.ts`](tests/spinner.spec.ts) | The animation sets (eight one-column frames, braille only on the activity line, eight different
|
|
341
|
+
| [`tests/spinner.spec.ts`](tests/spinner.spec.ts) | The animation sets (eight one-column frames, braille only on the activity line, eight different frames in each tool family, no glyph shared by two families, every family mixing flowers, blooms, and stars), tool-family mapping, activity mapping, the wall-clock frame step, the phase stagger, and the shimmer's sweep, rest, and disabled-palette output |
|
|
342
342
|
| [`tests/pace.spec.ts`](tests/pace.spec.ts) | The pacer's per-frame share, grapheme cuts, arrival order, flush, per-channel flush, and clear, and the row reveal's per-frame share, settle, and clamp |
|
|
343
343
|
| [`tests/stream-pace.spec.ts`](tests/stream-pace.spec.ts) | Paced reply text and tool arguments in stream order, thinking drawn when its block ends, the flush at stream end and before a logged event, reduced motion, and tool cards unrolling on the frame tick |
|
|
344
344
|
| [`tests/view-banner.spec.ts`](tests/view-banner.spec.ts) | The subagent-view line's trail, back target, width, and absence at the main session |
|
|
@@ -464,6 +464,7 @@ These limits describe the terminal surface as shipped; they are not a general CL
|
|
|
464
464
|
- **A frame that shrank leaves the top of the viewport out of reach** — the repaint window rises with the tallest frame written so far and never falls, so after read mode closes or the spinner leaves, a transient line is printed into the conversation instead of floating; it recovers as the conversation grows. The reader is unaffected: it draws on the terminal's other screen, where every row is its own.
|
|
465
465
|
- **The conversation is not drawn while the reader is up** — its screen is held off the terminal, so a turn that runs, a tool result that lands, and a notice that is printed are all drawn in one frame when the reader closes. The reader itself shows them as they arrive, because it re-reads the conversation on every drawing.
|
|
466
466
|
- **The reader shows source text** — an assistant reply reads as the Markdown the model wrote, not the rendering the transcript draws, so tables and headings read as their source; only a list item's leading `-`, `*`, or `+` is drawn as `•`.
|
|
467
|
+
- **A cold large turn needs a measurement pass** — before it can draw that turn's viewport, the reader counts every section to find the full scroll range; later paints reuse cached row counts and draw only visible sections, while changed content or a width it has not measured requires fresh measurements.
|
|
467
468
|
- **The panel lists residency, not the tree** — a child joins it while its session record is resident in this process, so a subagent run by an out-of-process provider, which owns no session here, never appears; `/subagents` remains the way to every durable descendant.
|
|
468
469
|
- **Residency is not work** — the listing's `activity: 'running'` says the child's record is resident, which is what the row's `resident` reports; whether the child is working is the separate `running` / `idle` word beside it, read from that child's Agent in this process.
|
|
469
470
|
- **Rows behind the overflow row are not selectable** — the panel draws at most six rows and `Up` / `Down` leave it at their ends; the children folded into `+<n> more` are reached through `/subagents`, which walks the complete descendant tree.
|
|
@@ -40,7 +40,7 @@ dsh tui --no-open # print sign-in URLs without opening a
|
|
|
40
40
|
|
|
41
41
|
### 屏幕布局
|
|
42
42
|
|
|
43
|
-
每一帧都从终端左右两边各内缩一列绘制,因此对话记录与停靠的界面组件都不会有任何一行贴到两侧边缘。页眉在标题生成或设置后以标题命名会话,并在旁边显示 id。对话记录在终端自身的回滚区中增长:你的提示以 `❯` 开头,启用颜色时铺在一条通栏的着色背景带上(附件列在其下),而在已有模型步骤的轮次中被引导或注入的提示以暗色文字铺在更深的背景带上,退到它所加入的轮次之后,assistant 的推理以暗色斜体显示在 `✻ Thinking` 标题之下、Markdown
|
|
43
|
+
每一帧都从终端左右两边各内缩一列绘制,因此对话记录与停靠的界面组件都不会有任何一行贴到两侧边缘。页眉在标题生成或设置后以标题命名会话,并在旁边显示 id。对话记录在终端自身的回滚区中增长:你的提示以 `❯` 开头,启用颜色时铺在一条通栏的着色背景带上(附件列在其下),而在已有模型步骤的轮次中被引导或注入的提示以暗色文字铺在更深的背景带上,退到它所加入的轮次之后,assistant 的推理以暗色斜体显示在 `✻ Thinking` 标题之下、Markdown 回复上方,每次工具调用是一张卡片,含状态符号(运行中为按工具族混合花朵、绽放与星形的黄色旋转指示、完成为绿色 `◆`、失败为红色 `◆`)、粗体工具名、以紫色显示的呈现器标题(去掉呈现器重复写出的工具名,因此 `Read src/app.ts` 画成 `read src/app.ts`;shell 卡片取该次调用自身的摘要,绝不取命令,命令只由卡片的 `$` 行绘制一次)、运行中的一行加载标记,以及折叠到 `toolPreviewLines` 行的正文。卡片在模型开始调用时出现,先画标题,其下各行在约 `toolRevealFrames` 帧内逐行展开;工具回答后,结果行也以同样方式展开并浮出。`edit` 或 `write` diff 卡片把每一新增行整行铺满绿色背景、每一删除行整行铺满红色背景,删除的旧行在上、新增的新行在下,并把 `+` 与 `-` 画成各自更亮的颜色,使符号从填充色中读得出来;折行续行只延续填充色,不再画第二个符号;工具应用变更后,已应用的 hunk 取代调用时携带的 diff,并在每行前标出文件行号,新增行与未变行用新文件的行号,删除行用旧文件的行号。`subagent` 与 `subagent_*` 卡片折叠时只画一行,行首是同样的状态符号,运行中旋转,交给后台的子 agent 则为暗色 `◆`:运行中显示工具名、任务描述,以及调用指定的模型与 `background`;结束后描述变暗,右端标出 `[done]`,交给后台的子 agent 标 `[started]`,失败标 `[failed]`。在被标记的行上按 `Space`,或按 `Ctrl+O`,即可展开完整卡片。非空的系统提示词与每一条注入的上下文——instructions、catalogs、snapshots、notices、relays 与 recalls——绘制为暗色 `⬡` 标题,其下是面向模型文本的前 `contextPreviewLines` 行,快照的每一份贡献具名写在其自身各行上方;空的系统提示词则省略。一次压缩(compaction)——无论来自 `/compact` 还是自动触发——绘制为一个上下文块,其强调色 `⬡` 标题写明来源,以及被替换内容与摘要各自的估算 token 数,如 `context compacted by /compact · ~6.4k → ~0.9k tokens` 或 `context compacted automatically in turn 12 · …`;标题之下,`compressed` 写明摘要替换的条目数与轮次、其中的每条提示(最多五条,每条一行),以及回复、按工具分列的工具结果、上下文块与更早摘要的计数,`preserved` 写明系统提示词以及之后每一轮仍逐字保留的内容;这两部分始终绘出,其下的摘要则像其他上下文块一样折叠到 `contextPreviewLines` 行。检查点消息本身不再重复绘制。日志以错误结束的 `/compact` 尝试会在会话重绘时打印 `/compact failed · <error>`,因为实时运行的命令会自己报告失败;失败的自动尝试不改变任何内容,也不绘制任何内容,与浏览器一致。承载的行数多于其绘制行数的工具卡片与上下文块,结尾都是同一条标记:该块被对话记录的焦点持有时为 `… <n> more rows · Space expands`,否则为 `… <n> more rows · Ctrl+O expands`。回复中的围栏代码、`read` 卡片的文件行、`edit` 或 `write` diff 卡片的变更行——其中 `+` 新增项标为绿色、`-` 删除项标为红色,符号后的源文本按文件扩展名对应的语言绘制——以及 `!` / `!!` 草稿、用户 shell 的 `$ command` 行(其 `$` 用强调色)与终端工具卡片中的 shell 命令,都使用语法高亮,配色主题由终端自身的背景在明暗之间选定;某种语言的语法会在第一个用到它的代码块出现时才加载,因此该块先以纯文本绘制、随后被重绘为彩色;这里没有对应语法的语言、既不支持 24 位也不支持 256 色的终端,以及 `codeHighlight: false`,都会把每个代码块画成纯文本。标题与列表标记使用暖橙色,链接与行内代码使用紫色,引用保持暗色。流式推理、回复文字与工具参数到达时先入队,再按帧(默认约每秒 60 帧)分批绘制,因此一次网络突发会分摊到 `streamPaceFrames` 帧里,而不是一次落下,队列也从不拖慢模型调用。流式回复文字会淡入,而且每个词各走各的时钟:一个词以接近终端背景色的亮度出现,并经 `streamFadeSteps` 次、每次 `streamFadeStepMs` 提亮到它最终稳定的颜色,因此流得更快只会留下更长的一串正在提亮的词,而不会更暗。流式推理与工具卡片在同一时长内浮出:它们以抬高的颜色出现,再退回到稳定时的暗色斜体或调色板颜色;从持久化历史重绘的卡片不带淡入,已经稳定下来的文字也不会再被调暗。
|
|
44
44
|
|
|
45
45
|
对话记录下方依次是 agent 工作时的旋转指示——`thinking`、`writing`、`calling <tool>`、重试或 `compacting`,各自旁边是它自己的盲文动画:环绕的点、逐行填满的线、绕行的单点、滑动的横条,或沉降又回满的方格——标签单词上带着一道扫过的高亮,并带着当前这次模型调用的实时 `↑` 发送与 `↓` 接收 token,以紧凑的 `k` / `M` / `B` 计数,随数据流入增长,并在提供方的 usage 块到达后落定——任何打开的提示、有消息等待时的带框 follow-ups 列表、本轮 todo 与最新后代行的 activity board(有行时才绘制;已完成内容划掉,状态符号以绿色、黄色或青色显示,进行中的那一项为旋转指示,todo 最多四行其后 `+<n> more`,不是焦点区域,在 `turn/end` 与 bind 时消失)、编辑器、子 agent 面板,以及页脚:未聚焦时是一行由两端向内构建的状态栏。编辑器的光标是终端自身的闪烁竖条:应用在启动时请求这一形状,退出时把你的默认形状还回去,而状态栏或面板持有键盘期间完全不绘制光标。模型与入口按键带有颜色;进行中的轮次、上下文与 todo 信息分别使用黄色、紫色与绿色。模型锚定在左端且永不丢弃——超过 20 列时退回到裸模型名,只有仍然过宽的标签才会被省略号截断——离开编辑器的那两个按键则锚定在右端,写作 `Shift+↑ read · Shift+↓ status`,有 follow-ups 等待时则写作 `Shift+↑ select · Shift+↓ status`,窄到一定程度收缩为 `Shift+↑↓ nav`,不足 40 列时整体丢弃。关键事实填满剩下的中间部分,按状态栏顺序排列:推理强度(选择把推理交给模型时标为 `effort default`)、进行中轮次的已用时间、上下文窗口百分比、todo 计数,以及 workspace 路径(过长时以 `~` 与 `…/` 缩短)。中间部分容纳不下的每个分段都折进末尾的 `+N`,未聚焦行从不绘制的那些分段也一并折入——权限预设、token 用量(会话合计加上进行中的这次调用)、来自投影接缝的目标与计划模式标记,以及待发送附件的数量。模型请求重试以通知形式出现,与浏览器标记承载的事实相同。
|
|
46
46
|
|
|
@@ -48,11 +48,11 @@ dsh tui --no-open # print sign-in URLs without opening a
|
|
|
48
48
|
|
|
49
49
|
被聚焦的块也会在原地被标记:一条两列宽的标记条,在块的其他行旁为暗色、在被聚焦小节自身的行旁为强调色——若折叠把该小节整个留在了绘出的行之外,强调色便落在那一行折叠标记旁——标记绘制期间该块的内容按窄两列换行。这个标记只能落在 pi-tui 仍以差分方式重绘的行上——也就是它写出的上一帧的最后 `rows` 行,这条边界只会被更高的帧抬高、不会被更矮的帧降低——因为改动其上方的任何一行都会清空终端的回滚区。已经滚过该位置的块不带标记条,检视面板的标题追加 ` · off screen`,其边框线也改画暗色而非强调色,于是边框本身就报告出标记无法画在该块所在之处。淡入中的各行抵达这一边界时,会在写出它们的同一帧内被结算为最终颜色,因此回滚区里不会留下暗色文字。
|
|
50
50
|
|
|
51
|
-
在任一区域按 `Ctrl+G`、在对话记录的某个小节上按 `Enter`,以及 `/turns`,都会整屏读出对话记录,一次读一个轮次。阅读器并不画在对话记录之上,而是接管终端:它运行在终端的备用屏幕(alternate screen,分页器所用的那块屏幕)上,只要它还开着,对话记录自己的屏幕就被挡在终端之外。因此它不占用对话记录的任何一行,让重绘边界保持原位,在任何终端上都铺满整屏,也不可能把自己的任何一行留在你的回滚缓冲里;关闭它会把对话记录原样恢复成你离开时的样子,期间落地的内容则绘制在其下方。轮次面板滚过首行后,紧凑的着色条会替代顶边线,将当前用户提示合并为一行,并按终端宽度以省略号截断;滚回顶部时恢复 ` ● READER ` 徽标与 `turn <n> of <total>`。在这个固定页眉之下,正在阅读的轮次占据屏幕的大部分,左侧是一列窄的轮次列表,每个正文行都以右边框收口。列表每个轮次一行——正在阅读的轮次带 `❯`、轮次编号、其提示的首行,以及右对齐的标记 `⬡<n>` 注入上下文块数、`✻` 有推理、`¶` 有回复、`◆<n>` 工具调用数——并在正在阅读的那个轮次之下列出它的各小节,`▸` 落在阅读器所持的那一个上,因此任何轮次自己的上下文都只有一步之遥。轮次面板用对话记录自己的符号把该轮次从头到尾滚动读出,相邻两个小节之间隔一空行:提示以 `❯` 开头、铺在着色背景带上,其后是暗色斜体的 `✻ Thinking`、按对话记录的方式渲染 Markdown(标题、强调、行内代码,以及带语法着色的围栏代码)但无序列表符号画成 `•` 的 `¶ Reply`、工具调用的 `◆`(结果落地后为绿色、运行中为黄色)连同粗体工具名与链接蓝色的标题、缩进在调用之下的 `⎿ Result`(diff 的 `+ ` 行为绿色、`- ` 行为红色,各铺在着色背景带上;`read` 卡片的文件行带语法着色),以及带注入上下文标签的 `⬡`。顶行所属的小节带强调色,并在每一行带上对话记录走查所用的 `┃` 标记条,无色时也能辨认。背景带由终端报告的背景色混合而成,因此没有报告背景色、或不绘制颜色的终端画出的面板不带背景带。那里的行从不折叠:阅读器正是读取系统提示词、注入快照、长回复或工具结果完整面向模型文本的地方。`Right` 打开被选中的轮次,`Left` 回到轮次列表;在列表中 `Up` / `Down` 每次步进一个小节并跨入相邻的轮次,而 `PageUp` / `PageDown` 步进整个轮次;在轮次面板中它们逐行滚动,而顶行所属的小节就是阅读器关闭后对话记录恢复到的位置。`/` 在按键提示的位置打开一行查询,把轮次列表收窄到提示或文本匹配的轮次,读数写作 `<kept>/<total> turns`;其余时候提示边线在右端报告 `turn 3/12 · row 14/212`,任何宽度都不会丢弃它。列数不足 `readerMinColumns` 时正文一次只画一个面板——轮次列表,或由 `Right` 从中打开的那个轮次——不足 24 列或 8 行时只画一行暗色文字 `terminal too small for the reader (needs 24×8)`,并且只应答 `Esc`、`Ctrl+G` 与 `Ctrl+C
|
|
51
|
+
在任一区域按 `Ctrl+G`、在对话记录的某个小节上按 `Enter`,以及 `/turns`,都会整屏读出对话记录,一次读一个轮次。阅读器并不画在对话记录之上,而是接管终端:它运行在终端的备用屏幕(alternate screen,分页器所用的那块屏幕)上,只要它还开着,对话记录自己的屏幕就被挡在终端之外。因此它不占用对话记录的任何一行,让重绘边界保持原位,在任何终端上都铺满整屏,也不可能把自己的任何一行留在你的回滚缓冲里;关闭它会把对话记录原样恢复成你离开时的样子,期间落地的内容则绘制在其下方。轮次面板滚过首行后,紧凑的着色条会替代顶边线,将当前用户提示合并为一行,并按终端宽度以省略号截断;滚回顶部时恢复 ` ● READER ` 徽标与 `turn <n> of <total>`。在这个固定页眉之下,正在阅读的轮次占据屏幕的大部分,左侧是一列窄的轮次列表,每个正文行都以右边框收口。列表每个轮次一行——正在阅读的轮次带 `❯`、轮次编号、其提示的首行,以及右对齐的标记 `⬡<n>` 注入上下文块数、`✻` 有推理、`¶` 有回复、`◆<n>` 工具调用数——并在正在阅读的那个轮次之下列出它的各小节,`▸` 落在阅读器所持的那一个上,因此任何轮次自己的上下文都只有一步之遥。轮次面板用对话记录自己的符号把该轮次从头到尾滚动读出,相邻两个小节之间隔一空行:提示以 `❯` 开头、铺在着色背景带上,其后是暗色斜体的 `✻ Thinking`、按对话记录的方式渲染 Markdown(标题、强调、行内代码,以及带语法着色的围栏代码)但无序列表符号画成 `•` 的 `¶ Reply`、工具调用的 `◆`(结果落地后为绿色、运行中为黄色)连同粗体工具名与链接蓝色的标题、缩进在调用之下的 `⎿ Result`(diff 的 `+ ` 行为绿色、`- ` 行为红色,各铺在着色背景带上;`read` 卡片的文件行带语法着色),以及带注入上下文标签的 `⬡`。顶行所属的小节带强调色,并在每一行带上对话记录走查所用的 `┃` 标记条,无色时也能辨认。背景带由终端报告的背景色混合而成,因此没有报告背景色、或不绘制颜色的终端画出的面板不带背景带。那里的行从不折叠:阅读器正是读取系统提示词、注入快照、长回复或工具结果完整面向模型文本的地方。`Right` 打开被选中的轮次,`Left` 回到轮次列表;在列表中 `Up` / `Down` 每次步进一个小节并跨入相邻的轮次,而 `PageUp` / `PageDown` 步进整个轮次;在轮次面板中它们逐行滚动,而顶行所属的小节就是阅读器关闭后对话记录恢复到的位置。`/` 在按键提示的位置打开一行查询,把轮次列表收窄到提示或文本匹配的轮次,读数写作 `<kept>/<total> turns`;其余时候提示边线在右端报告 `turn 3/12 · row 14/212`,任何宽度都不会丢弃它。列数不足 `readerMinColumns` 时正文一次只画一个面板——轮次列表,或由 `Right` 从中打开的那个轮次——不足 24 列或 8 行时只画一行暗色文字 `terminal too small for the reader (needs 24×8)`,并且只应答 `Esc`、`Ctrl+G` 与 `Ctrl+C`。它每一帧都会重新读取对话记录,只绘制视区内的小节,并在内容、宽度和样式与缓存匹配时复用已有行,因此长轮次在会话运行期间仍能保持响应;流式回复会在其中增长、落地的工具结果会出现、新轮次会加入轮次列表;重新换行时它按正在阅读的小节重新锚定,而不是按行号;审批或提问出现时它在该提示存续期间让位;切换会话会让它随通知 `the transcript changed · reader closed` 一同关闭。
|
|
52
52
|
|
|
53
53
|
临时按键反馈以一行暗色带框文字浮在视口右上角:`press Esc again to stop turn <n>`、`press Ctrl+C again to quit` 与 `nothing in the transcript to read yet`。它以全亮度保持 `toastMs`、随后淡出,不占用对话记录自身的任何一行,也不接管键盘;渲染器已无法重绘视口顶部时,它改为打印进对话记录,从而被保留下来。值得留存的事实则是对话记录中的通知:`stopping the turn…`、`wait for the session switch to finish`、`above the repaint window · opened in the reader` 以及每一条命令结果。
|
|
54
54
|
|
|
55
|
-
动效不止于流式文本:每种抬亮与浮出都乘着同一个淡入时钟,每种工作动画则乘着同一个旋转时钟。键盘落到对话记录、子 agent 面板或状态栏上时,会把该表层的边框、其徽标与它持有的标记抬亮十二拍(每拍一个 `streamFadeStepMs`);对话记录走查中的一步把新聚焦的标记条抬亮六拍;状态栏上的一步把所选标签抬亮八拍。活动板上新出现或状态已变的 todo 行,以及被替换的后代行,也在淡入时钟上像工具卡片一样浮出;重放与 `reducedMotion` 则画出已稳定的活动板。每种工作动画都有八帧,由时钟决定当前画哪一帧,并在周期为 `spinnerMs`(默认 100 ms
|
|
55
|
+
动效不止于流式文本:每种抬亮与浮出都乘着同一个淡入时钟,每种工作动画则乘着同一个旋转时钟。键盘落到对话记录、子 agent 面板或状态栏上时,会把该表层的边框、其徽标与它持有的标记抬亮十二拍(每拍一个 `streamFadeStepMs`);对话记录走查中的一步把新聚焦的标记条抬亮六拍;状态栏上的一步把所选标签抬亮八拍。活动板上新出现或状态已变的 todo 行,以及被替换的后代行,也在淡入时钟上像工具卡片一样浮出;重放与 `reducedMotion` 则画出已稳定的活动板。每种工作动画都有八帧,由时钟决定当前画哪一帧,并在周期为 `spinnerMs`(默认 100 ms)的旋转时钟上步进,因此一切旋转中的指示都在同一时刻、以同一速度前进,标签变化也不打乱节拍,而该时钟只在其中之一被绘制时才启动。只有工作时的旋转指示画盲文,其动画随状态而变,因此它从不与卡片的符号重复。运行中的工具卡片与折叠的子 agent 行画该工具族的动画,进行中的 todo 行画规划族的动画。这些动画都把花朵、绽放与星形混在同一组帧里,族与族之间不共用符号,同一族的每一帧也各不相同,因此同一族里相位不同的两个指示在任何时刻都不会画出相同的符号。某个指示开始旋转时若同族已有其他指示在转,它会取一个相位,使它在任何时刻都不与它们画出相同的符号,同族同时旋转的不超过八个时都能做到;超过八个后,它取同族中被持有指示最少的那个相位。标题行已滚动到渲染器可重绘区域之上的卡片自此画出静态的 `◆`,而 `reducedMotion` 把它们全部画成静止的——卡片的 `◆`、todo 的 `▸`,以及不带符号、只剩标签的工作旋转指示——且不启动旋转时钟。阅读器在自己占据的屏幕上一次性完整出现,而它新显示的内容会像工具卡片一样在淡入时钟上浮出:打开它与显示另一个轮次时浮出整个轮次,在列表中步进到同一轮次的另一个小节时浮出该小节,历时 `streamFadeSteps` 拍;滚动、实时增长与重新换行都不浮出任何内容。任何抬亮都不改变任何东西绘制的行数,而 `reducedMotion`——与 `NO_COLOR`、被禁用的调色板和 `TERM=dumb` 一样——会关闭每一种抬亮与浮出,并且不为它们安排任何重绘。
|
|
56
56
|
|
|
57
57
|
子 agent 面板在绑定会话之下有常驻的子 agent 会话、或列表中带有无法读取的候选者时绘制。未聚焦时它是一行摘要,含列出的数量与第一个子会话的关键标签;聚焦时其标题统计所列出的条目,并以它所应答的按键 `↑↓ children · Enter opens · Tab regions · Esc input` 结尾,在容纳不下时收缩为 `↑↓ children · Esc input`,再收缩为 `Esc input`;每一行给出该子会话的层级缩进、其标签或 id、其模式(`one-shot` 或 `continuable`)、`resident`、其 agent 处于 `running` 还是 `idle`、其已用时间——进行中轮次的用时,否则是已结束轮次的合计——以及其 token 用量,以本进程对该子会话的可见程度与已组合的投影所能提供的为限。最多绘制六行,其下是 `+<n> more · /subagents lists them all`;列表无法解读的候选者绘制为 `unreadable: <reason>` 且打不开任何页面,读取失败的列表则保留上一次成功读取产生的各行,并在其下写出 `listing failed: <reason>`。面板随其最后一行一同消失。在可读的行上按 `Enter` 或 `Right`,与在可读的 `/subagents` 行上按 `Enter` 一样,会把该子会话作为子 agent 视图打开:终端在已绑定会话的位置绘制子会话——其对话记录、实时流、对话记录走查、检查器、阅读器以及它自己的子 agent 面板——而进入它之前的会话保持打开。常驻的子会话被实时读取,其他子会话则被恢复。条目的详情行作为通知打印在 `subagent <label> · Ctrl+P or /parent returns to session <id>` 之下,页眉以 `◆ subagent view ›`(每进入一层一个 `›`)与 `· Ctrl+P returns` 结尾,只要视图打开,输入框正上方就有一条着色行写着 `◆ subagent view · main › <label> · Ctrl+P back to <parent>`,无论对话记录滚动到哪里,`Ctrl+P` 或 `/parent` 返回上一层,并按父会话此刻的日志重绘它。只要这条链中任一会话正在运行轮次,`/new`、`/clear`、`/resume`、`/sessions` 与 `/fork` 都会拒绝,否则它们会随所离开的会话释放每一个视图;退出会释放每一个视图并保存根会话。
|
|
58
58
|
|
|
@@ -306,7 +306,7 @@ runner 等待完整应用就绪(`ctx.get('loader')?.await()`),并在核心
|
|
|
306
306
|
| [`src/fade.ts`](src/fade.ts) | 流式文本淡入与浮出:基于挂钟的尾部追踪器、块时钟与注册表、淡入亮度级别、浮出混合,以及对已渲染行的重新着色 |
|
|
307
307
|
| [`src/pace.ts`](src/pace.ts) | 流式节拍器:助手流与对话记录之间的有序队列,每帧释放积压的一份;以及以同样方式展开工具卡片的逐行展开器 |
|
|
308
308
|
| [`src/motion.ts`](src/motion.ts) | 边框装饰的动效时钟与其调用点所绘的三级抬亮:落键、小节步进与状态栏走查 |
|
|
309
|
-
| [`src/spinner.ts`](src/spinner.ts) |
|
|
309
|
+
| [`src/spinner.ts`](src/spinner.ts) | 工作动画:活动行按状态区分的盲文动画、每个工具族用于运行中卡片、子 agent 行与进行中 todo 的花朵、绽放与星形动画、哪个工具画哪一族、每个指示在某一时刻画出的帧、让同一族的指示彼此错开的相位,以及活动标签所带的扫光 |
|
|
310
310
|
| [`src/prompts.ts`](src/prompts.ts) | 审批、提问、选择器与只读详情提示以及模态队列 |
|
|
311
311
|
| [`src/toast.ts`](src/toast.ts) | 临时按键提示行:其浮层、其时钟,以及它承载的各条文字 |
|
|
312
312
|
| [`src/pinned-prompt.ts`](src/pinned-prompt.ts) | 阅读器在轮次面板滚动后固定在顶部的提示行:符号、合并为一行的提示,以及背景带 |
|
|
@@ -338,7 +338,7 @@ runner 等待完整应用就绪(`ctx.get('loader')?.await()`),并在核心
|
|
|
338
338
|
| [`tests/editor.spec.ts`](tests/editor.spec.ts) | 对照 pi-tui 编辑器行为验证终端光标、`Shift+Left` / `Shift+Right` 词导航,以及 `!` / `!!` 草稿着色 |
|
|
339
339
|
| [`tests/frame.spec.ts`](tests/frame.spec.ts) | 边框线、徽标、正文行,以及某个宽度容纳得下的按键提示 |
|
|
340
340
|
| [`tests/motion.spec.ts`](tests/motion.spec.ts) | 动效时钟的各个级别、它对重绘的要求,以及每一级所绘的抬亮 |
|
|
341
|
-
| [`tests/spinner.spec.ts`](tests/spinner.spec.ts) |
|
|
341
|
+
| [`tests/spinner.spec.ts`](tests/spinner.spec.ts) | 各组动画(八个单列帧、盲文只用于活动行、每个工具族的八帧互不相同、两族不共用符号、每个族都混合花朵、绽放与星形)、工具族的映射、活动的映射、按挂钟的帧步进、相位错开,以及扫光的扫过、停顿与调色板禁用时的输出 |
|
|
342
342
|
| [`tests/pace.spec.ts`](tests/pace.spec.ts) | 节拍器的每帧份额、按字素切分、到达顺序、flush、按通道 flush 与 clear,以及逐行展开器的每帧份额、settle 与 clamp |
|
|
343
343
|
| [`tests/stream-pace.spec.ts`](tests/stream-pace.spec.ts) | 按流顺序分批绘制的回复文字与工具参数,思考块结束时画出思考,流结束时与已记录事件前的 flush,reducedMotion,以及在帧节拍上展开的工具卡片 |
|
|
344
344
|
| [`tests/view-banner.spec.ts`](tests/view-banner.spec.ts) | 子 agent 视图行的路径、返回目标、宽度,以及在主会话时不绘制 |
|
|
@@ -464,6 +464,7 @@ The user moved this call to the background. It keeps running as job <id>. You ar
|
|
|
464
464
|
- **缩过的帧会让视口顶部脱离可及范围**——重绘窗口随迄今写出的最高一帧上升且不会回落,因此在读取模式关闭或旋转指示离开之后,临时提示行改为打印进对话记录;它会随对话记录增长而恢复。阅读器不受影响:它画在终端的另一块屏幕上,那里的每一行都归它所有。
|
|
465
465
|
- **阅读器开着时不绘制对话记录**——对话记录的屏幕被挡在终端之外,因此期间跑完的轮次、落地的工具结果与打印的通知,都在阅读器关闭时的同一帧里绘出。阅读器自身仍会随它们到达而显示,因为它在每一次绘制时都重新读取对话记录。
|
|
466
466
|
- **阅读器显示的是源文本**——assistant 回复读到的是模型写下的 Markdown,而不是对话记录绘制出的渲染结果,因此表格与标题以源码形式呈现;只有列表项开头的 `-`、`*` 或 `+` 画成 `•`。
|
|
467
|
+
- **首次读取大型轮次仍需测量**——在绘制该轮次视口前,阅读器会统计每个小节以得出完整滚动范围;后续绘制会复用缓存的行数并只绘制可见小节,但内容变化或尚未测量过的宽度仍需重新测量。
|
|
467
468
|
- **面板列出的是常驻,而不是整棵树**——子会话在其会话记录常驻于本进程期间加入面板,因此由进程外提供方运行、在此没有自己会话的子 agent 永远不会出现;`/subagents` 仍是到达每一个持久后代的途径。
|
|
468
469
|
- **常驻不等于正在工作**——列表的 `activity: 'running'` 表示该子会话的记录常驻,这正是行内 `resident` 所报告的内容;子会话是否在工作则是它旁边单独的 `running` / `idle`,读自本进程中该子会话的 Agent。
|
|
469
470
|
- **溢出行之后的各行不可选中**——面板最多绘制六行,`Up` / `Down` 在其两端会离开面板;被折叠进 `+<n> more` 的子会话通过 `/subagents` 到达,后者遍历完整的后代树。
|