src-misc/system-prompt.txt masih memakai framing lama "Zesdex Corp
CEO / 5 divisi" dan mereferensikan tool run_company_pipeline yang
sudah tidak ada, tertinggal saat commit 25f084f merombak arsitektur
ke hive-mind (CLAUDE.md, README.md, dan implementasi workflow_run
sudah diupdate saat itu, tapi file prompt utama ini terlewat).
Akibatnya AI membalas dengan persona CEO/company, bukan Core
Intelligence/hive-mind seperti didokumentasikan di README.md.
- system-prompt.txt: tulis ulang total ke model Core Intelligence /
cognitive cycle plan / processing node / access tier, konsisten
dengan CLAUDE.md dan tool hive_mind yang sebenarnya.
- system-tools.txt: tambah entri hive_mind dan read_findings yang
sebelumnya tidak ada sama sekali di daftar tool.
- workflow.rs: perbaiki sisa teks "divisions/subagents" di deskripsi
tool read_findings jadi "nodes/subagents".
Diverifikasi live: AI sekarang memperkenalkan diri sebagai "Core
Intelligence of Zesdex ... modeled as a hive-mind" alih-alih CEO.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Overlay Todo sebelumnya menampilkan jumlah pesan yang tidak relevan,
bukan isi task list. Overlay Usage sekarang pakai compute_usage_summary
yang sama dengan widget sidebar (DRY, angka konsisten).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Baris pertama widget Usage sebelumnya bisa melebihi lebar kolom sidebar
(30 kolom) pada sesi dengan jumlah token besar, menyebabkan teks
terpotong diam-diam tanpa wrap. Sekarang dipecah jadi beberapa baris
pendek yang aman di lebar berapa pun yang realistis.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Sidebar kanan permanen (Workflow/Tasks/Usage) menggantikan panel todo
ad-hoc yang lama. Widget baca state yang sudah ada, tidak ada perubahan
skema AppStateRest.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Overlay Todo dan Usage sebelumnya tidak punya trigger sama sekali di
jalur interaksi normal (cuma bisa lewat restore snapshot sesi) --
sekarang mengikuti pola /workflow yang sudah ada.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Konten Tool sebelumnya didorong sebagai satu Span tanpa split newline,
sehingga output tool multi-baris (stdout bash, hasil grep, diff)
tampil sebagai karakter \n literal alih-alih baris terpisah. Sekarang
mengikuti pola render_markdown + split_spans_into_lines yang sama
dengan role lain.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
markdown.rs juga disesuaikan: indentasi paragraf/heading bawaannya
dilepas supaya tidak bentrok dengan indent PREFIX_WIDTH di chat.rs
(baris pertama vs baris wrap lanjutan jadi sejajar).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
unnested_or_patterns, collapsible_if, items_after_statements, dan
case_sensitive_file_extension_comparisons di auto.rs, engine.rs, dan
actions/mod.rs.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Spec brainstorming untuk rombak total src/view + src/controller: layout
Multi-Pane Dashboard dengan sidebar Workflow/Tasks/Usage, palet Tokyo
Night, dan format chat inline rapat.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Pesan sistem bertanda [Hive-Mind Consensus] hanya di-push ke variabel lokal
run_agent_turn dan diarsipkan ke SQLite, tidak pernah masuk ke
rt.messages lewat TurnEvent — sehingga hive_mind_already_ran selalu
memindai daftar pesan yang kosong dan gerbang "converge sekali per sesi"
tidak pernah aktif. Tambahkan SessionRuntime.hive_mind_converged yang
diset dari event TurnEvent::SystemNote { kind: "hive_mind_converged" }
setelah konvergensi selesai, disalurkan lewat TurnCtx, dan dijadikan
sinyal utama di run_agent_turn (pemindaian pesan lama tetap sebagai
fallback defensif).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
RunningGuard resets TEST_GEN_RUNNING/ARCH_REVIEW_RUNNING/SECURITY_REVIEW_RUNNING
via Drop so a subagent panic can no longer wedge that review kind disabled
for the rest of the process. Doc comments on run_subagent_with_retry and the
three spawn_background_* / spawn_all_background functions now describe the
abort_flag and overlap-guard behavior added in Task 7.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
- Introduced a new hive-mind architecture that allows the Core Intelligence to issue directives to anonymous processing nodes.
- Each node executes its directive and merges output into a collective state, visible to all nodes in real-time.
- Added support for dynamic cognitive cycles, enabling flexible task management.
- Implemented documentation generation for hive-mind runs, ensuring a durable record of decisions and actions.
- Refactored existing company pipeline tools to align with the new hive-mind structure, replacing division-specific prompts with a more generalized approach.
- Updated workflow rendering to accommodate hive-mind nodes and their system-assigned designations.
- Enhanced error handling and validation for cognitive cycle plans.
- Updated README.md to reflect the addition of 3 new built-in tools, bringing the total to 37.
- Revised architecture documentation to indicate the increase in tool count.
- Enhanced backend documentation with updated line counts for various modules.
- Modified data documentation to change edit log format from JSON to JSONL.
- Updated dependencies documentation to reflect version upgrades for several crates.
- Improved prompts for auto-reviewer, division implementer, planner, tester, and quality reviewer to enforce stricter coding standards regarding linter bypasses.
- Refactored code in various modules to improve clarity and performance, including updates to error handling and tool execution logic.
- Added comprehensive tests for IPC frame serialization and deserialization.
- Updated `crossterm` from version 0.28 to 0.29.
- Upgraded `reqwest` from version 0.12 to 0.13 and added "form" feature.
- Bumped `serde_yaml_ng` from version 0.9 to 0.10.
- Increased `dirs` version from 5 to 6.
- Updated `rusqlite` from version 0.32 to 0.40.
- Upgraded `infer` from version 0.16 to 0.19.
- Bumped `sha2` from version 0.10 to 0.11.
- Updated `rmcp` from version 1.8 to 2.2.
- Refactored SHA256 hash encoding in `rewind.rs`, `mod.rs`, and `rest.rs` to use `hex::encode` instead of formatting with `{:x}` for better clarity and consistency.
- Updated status bar rendering in `status.rs` to feature a segmented design with clear visual segments for app name, status, and metadata.
- Refined color theme in `theme.rs` to adopt a modern dark palette with neon accents, improving visual hierarchy and readability.
- Revamped workflow panel in `workflow.rs` to display agent statuses as compact cards with state badges, enhancing clarity and user experience.
- Improved overall styling consistency across components, ensuring a cohesive look and feel throughout the TUI.
- Introduced prompts for various subagent roles: architecture reviewer, code quality reviewer, documentation maintainer, implementation team, testing team, and security reviewer.
- Implemented the auto-subagent orchestration in `auto.rs` to manage inline and background reviews.
- Created a division structure in `division.rs` to define roles and responsibilities for each subagent.
- Developed a company workflow orchestrator in `company.rs` to run the complete division pipeline, consolidating findings and generating executive summaries.
- Added logic to determine whether to run a full or quick pipeline based on request complexity.
- Added new workflow commands: `/workflow` to open the workflow panel and `/workflow run <prompt>` to execute workflows.
- Implemented `spawn_agents` and `spawn_pipeline` tools for parallel and sequential task execution, respectively.
- Enhanced workflow engine to handle real-time agent status updates and display in the UI.
- Updated workflow panel to show agent statuses, findings count, and session counters.
- Refactored existing code to integrate new workflow functionalities and improve overall structure.
When ~/.claude/settings.json is found, registers model roles for:
- claude-opus-4-8 (200k context, 8192 max tokens)
- claude-sonnet-5
- claude-haiku-4-5-20251001
Sets claude-opus-4-8 as default model when no custom default exists.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Reads ANTHROPIC_BASE_URL and ANTHROPIC_API_KEY from Claude Code's
settings.json and registers them as a 'claude' provider. Automatically
sets it as the default provider when no custom default is configured.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Auto-install improvements:
- Add curl-based download helpers for rust-analyzer and jdtls
- Add progress callback threaded through all provision stages
- Check ~/.local/share/zesdex/lsp/ for previous downloads so install
only runs once, not on every startup
- Increase run_command timeout to 180s for large downloads
- Add cargo/pacman/brew fallback tiers for Rust
- Add pacman/download tiers for Java jdtls
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Added LSP auto-provisioning functionality to automatically install and connect language servers.
- Introduced `shutdown_lsp` method to cleanly shut down LSP servers on application exit.
- Enhanced `AppStateRest` to spawn a background thread for provisioning language servers.
- Updated `builtin_agents` to include new LSP-related agents.
- Modified settings to include options for LSP auto-provisioning and supported languages.
- Updated file editing and writing tools to notify LSP servers of changes.
- Enhanced LSP tools to support auto-detection of servers based on file extensions.
- Added utility functions for managing known file extensions and resolving server names.
- Created a new `provisioner` module to handle the provisioning logic for various language servers.
- Added LspClient for handling communication with LSP servers, including methods for initialization, notifications, and requests.
- Introduced LspManager to manage multiple LSP server connections, allowing for connection, disconnection, and retrieval of server capabilities.
- Created tools for connecting to LSP servers, retrieving diagnostics, hover information, code completion, definitions, and references.
- Enhanced UI rendering to display token usage and settings in the overlay.
- Updated status bar to show current token usage and selected provider/model.
- Added module-level documentation for memory tools (`remember`, `recall`, `forget`) to clarify their purpose.
- Improved documentation in `recall.rs` and `remember.rs` to describe the functionality and flow of memory entry operations.
- Updated `mod.rs` to include descriptions for the tool trait and execution context.
- Enhanced `plan.rs` with detailed comments on plan-mode signaling tools.
- Documented text search tools in `search.rs` to explain their functionality.
- Improved sequential-thinking tool documentation in `seqthink.rs`.
- Added safety filter documentation in `shell_filter` for credential and git operations.
- Enhanced utility tools documentation, including `cd`, `dir_cache_update`, and `todowrite`.
- Improved rendering documentation in view modules (`chat`, `markdown`, `status`, `workflow`) to clarify rendering flows and purposes.
- Replace all eprintln! with tracing::warn! to avoid TUI corruption
via stderr writes during alternate screen mode
- Route tracing output to ~/.local/share/zesdex/zesdex.log instead of
stderr by configuring tracing_subscriber with a Mutex<File> writer
- Add render_toasts() widget: floating notification stack at top-right
of the terminal, color-coded by severity (Info/Success/Warning/Error/
Lesson), auto-expires after 5s, max 4 visible
- Apply to 12 files: stream parser, MCP client, subagent engine,
state/rest, controller/input, app_config, provider, OAuth, agent_def,
dto/chat/tool
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add eprintln! logging before fallback values in 7 files where errors
were previously swallowed without visibility:
- [stream] malformed JSON chunk, missing usage tokens, missing tool call index
- [mcp] missing result fields, client builder failure, response body read error
- [subagent] missing API key in settings, all resolution paths exhausted
- [state] memory_dir no-parent, session_id missing, lock poisoned, store_base_dir
- [input] missing default_model for provider
- [session] corrupt agents.json parse failure
- [pkce] clock-before-epoch on random byte generation
All fallback values are preserved — this adds observability without
changing behavior for callers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Deleted the `security` module and its associated files, including `daemon.rs` and `install.rs`.
- Removed references to security features in various modules, including `mod.rs`, `mode/mod.rs`, and `input.rs`.
- Updated the `MiscState` struct to eliminate security-related fields.
- Adjusted the `apply_action` function to remove security action handling.
- Increased the maximum limits for tool-only turns and agent steps in `actions/mod.rs`.
- Modified the review prompt to exclude security checks.
- Cleaned up the `git_operator` and `shell` tools to remove catastrophic guard checks.
- Removed internet-related tools and their references from the tool module.
- Removed the Escalate variant from the Verdict enum and associated parsing logic.
- Cleaned up the EditorState struct by removing unused fields and methods.
- Simplified MCP connection functions by removing unnecessary disconnect and handle dismiss functions.
- Added tick count to MiscState for managing UI updates during processing.
- Updated chat and status views to display a spinner during AI processing using the tick count.
- Created a README.md file to document the project, its features, architecture, usage, key bindings, agent modes, security measures, installation instructions, and configuration details.
- Updated `scroll_up` and `scroll_down` methods to take an `amount` parameter for more flexible scrolling.
- Removed the `AgentMode` enum and related methods from the types module to simplify state management.
- Modified `AppStateRest` to remove the `mode` field and adjusted related logic.
- Enhanced `run_subagent` to build tool definitions and handle API key resolution from configuration.
- Updated command parsing to reflect changes in login handling.
- Removed onboarding overlays and related logic from input handling and rendering.
- Improved status bar to reflect connection status and agent readiness.
- Adjusted workflow panel rendering to simplify phase status display.
- Refactored edit log initialization to load from disk if available.
- Updated settings structure to use a HashMap for API keys.
- Enhanced error handling in LlmClient for authentication issues.
- Implemented `Edit` and `McpAdd` commands in the command parser and handler.
- Added a new `stream` module to the runtime for handling streaming events.
- Enhanced `SseParser` to parse usage information from SSE events.
- Introduced `ToolCallAccumulator` for tracking tool calls independently.
- Updated `AppStateRest` to include `app_config` and `MiscState` to track `effort_level` and `selected_index`.
- Modified `LlmClient` to support streaming responses with usage tracking.
- Improved error handling and retry logic in the streaming API calls.
- Added tests for new features and improved markdown rendering in the chat view.
- Removed unused structs and methods from `response.rs`, `usage.rs`, and `client.rs`.
- Simplified `Connection` handling in `conn.rs` to only support Unix sockets.
- Updated `IpcServer` to exclusively use Unix sockets and removed TCP handling.
- Cleaned up `editlog.rs` by removing loading and recent entry methods.
- Refactored `memory.rs` to eliminate unused functions related to lesson promotion and retrospective creation.
- Enhanced `search.rs` to support multiple search providers and improved error handling.
- Updated chat view logic to simplify message display and improve user experience.
- Removed deprecated modules and constants from various files to streamline the codebase.
- Removed unused modules and updated module paths for clarity.
- Added autocomplete functionality for command input in InputState.
- Updated AppStateRest to include a method for checking if a turn is in flight.
- Refactored subagent engine to use new API client structure.
- Changed default provider from "openrouter" to "zen" with updated API keys and models.
- Implemented tests for memory management and edit log functionalities.
- Enhanced error handling in API requests and improved response parsing.
- Updated UI components to reflect new API provider and status indicators.
- Add `zesdex_sec_daemon` module with main entry point for running the security daemon.
- Implement `TieredInstaller` for installing security tools from various sources (pip, binaries, gems).
- Create a newline-delimited JSON frame protocol for communication between the daemon and tools.
- Introduce a `ToolRegistry` for managing and dispatching tool executions.
- Add various tools including HTTP, SQLMap, Nuclei, and more with their respective execution logic.
- Establish health check and installation commands for tool management.
- Include prompts for classifier and quality reviewer to enhance code review and safety checks.
- Document the system's tools and guidelines for usage.
asepharyana