CHANGELOG.

No release notes provided.

Highlights

• i18n Fix (#109) — Language selection now applied to agents, prompts, and voice STT
• Approval Flow — Real WebSocket approval with 30min timeout (no more auto-approve stub)
• Conversation Tree — session-aware, persists conversation_id in session store
• Live Logs Tab — real-time monitoring in MonitoringScreen
• Invisible Backend — pythonw.exe, no console window
• ConnectionGuard — overlay when backend unreachable
• DB Migration — auto-copies ~/.jarvis/ data to ~/.cognithor/

All 6 GitHub Issues Fixed

• #104 first_run.py KeyError
• #105 model_post_init auto-swap override
• #106 upgrade corruption
• #107 connection-lost overlay
• #108 invisible backend + live logs
• #109 i18n language selection

Infrastructure

• Kanban sqlcipher cursor fix
• Evolution goal progress endpoint now reads from GoalManager
• Duplicate plan prevention (cleaned up 323 duplicates)
• Planner retries on action refusal

See CHANGELOG.md for full details.

Highlights

• Package rename complete: jarvis → cognithor (both CLI commands preserved)
• Release pipeline hardened: Pre-release smoke test validates built wheel on 2 OS before PyPI upload
• All 5 GitHub issues fixed: #104, #105, #106, #107, #108

Bug Fixes

• #104 first_run.py KeyError — config structure fixed from nested to flat
• #105 Model switching — model_post_init no longer overrides explicit llm_backend_type when API keys exist in keyring
• #106 Upgrade corruption — auto-migrate ~/.jarvis/ → ~/.cognithor/, keyring fallback, corrupt DB recovery
• #107 Frontend disconnect — ConnectionGuard overlay blocks UI when backend unreachable, 15s health polling
• #108 Invisible backend — pythonw.exe in --ui mode, Live Logs tab in MonitoringScreen

Infrastructure

• Environment variable overrides: JARVIS_* prefix supported for backward compatibility
• Version consistency check in CI (pyproject.toml vs init.py)
• scripts/prepare_release.py for local pre-release validation
• scripts/verify_readme_claims.py for README claims audit
• README: Pre-v1.0 Beta disclaimer, corrected numeric claims

New Tests

29 new tests: release smoke (4), env overrides (18), security regression (2), migration (5)

Full Changelog

See CHANGELOG.md for complete details.

Two Major Memory Subsystems

Both features from the Retrieval Extensions prompt are now complete.

CAG Layer — KV-Cache Preloading

• Deterministic prefix generation ensures LLM backends (Ollama, llama.cpp, vLLM) automatically reuse their internal KV-cache for Core Memory
• Content normalizer with BOM strip, whitespace collapse, sha256 hashing
• Cache store with atomic writes, JSON + txt sidecar
• Selector determines which memory tiers qualify (Core always, Semantic conditional)
• Metrics track prefix hit/miss rate
• PrefixCacheBuilder for any HTTP-based backend
• NativeLlamaCppBuilder for explicit state save/load (llama-cpp-python)
• 2 integration hooks: Gateway prepares prefix, Planner uses it instead of raw core_memory_text
• Config: opt-in (cag.enabled=false), auto-rebuild on change
• 71 tests

Hierarchical Document Reasoning — 4th Retrieval Channel

• 5 document parsers: Markdown, PDF, DOCX, HTML, Plaintext (German legal heuristics)
• Tree builder: hierarchy construction, content splitting, bottom-up LLM summaries
• SQLite tree store with CASCADE delete
• LLM node selector: top-down traversal with JSON parsing + regex fallback
• 4-channel score fusion: auto-normalizing weights (vector + BM25 + graph + hierarchical)
• 4 new MemoryManager methods, HierarchicalConfig
• 136 tests, 14 edge cases

Combined: 207 new tests, ~35 new files

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.88.0...v0.89.0

New Feature: 4th Retrieval Channel

Hierarchical Document Reasoning — index long structured documents (contracts, manuals, legal texts) as hierarchical trees and let the LLM navigate to relevant sections via top-down semantic reasoning. No vector embeddings needed.

Components

• 5 Document Parsers: Markdown (ATX+Setext), PDF (TOC+font-size), DOCX (heading styles), HTML (h1-h6+filtering), Plaintext (German legal heuristics)
• Tree Builder: Flat sections -> hierarchy -> content splitting -> bottom-up LLM summaries
• SQLite Tree Store: Same DB as existing indexer, CASCADE delete, transactional writes
• LLM Node Selector: Top-down traversal with JSON parsing + regex fallback
• 4-Channel Score Fusion: Auto-normalizing weights (vector + BM25 + graph + hierarchical)

Integration

• 4 new MemoryManager methods: index/remove/list/reindex hierarchical documents
• HierarchicalConfig in config.yaml (enabled by default, score_weight=0.25)
• Existing 3-channel search unchanged when no hierarchical docs exist

Quality

• 136 new tests (models, 5 parsers, store, builder, selector, retrieval, integration, 14 edge cases)
• All existing memory tests pass unchanged
• 14 edge cases covered (no headings, heading jumps, corrupt files, concurrent indexing, encoding issues, etc.)

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.87.2...v0.88.0

v0.87.2 — Zero Silent Failures + Install Overhaul

Silent Exception Audit

• 79 except Exception: pass blocks replaced with log.debug(..., exc_info=True) across 31 files
• Zero silent exception swallowing remaining in codebase

install.bat Overhaul (#100)

• PATH append (not overwrite), exact model matching, VRAM validation
• Network retry (3 attempts), disk space check, AMD GPU detection
• --silent mode, install log, COGNITHOR_OLLAMA_URL env var
• All 13 audit points addressed

install.sh Overhaul (#101)

• set -eo pipefail (removed -u), retry() for all network ops
• Multi-distro: apt-get, dnf, pacman, brew
• Portable: no GNU timeout, nproc/sysctl fallback, POSIX sort
• --silent mode, disk space check, install log
• macOS compatible (sysctl, no systemd assumption)

Bugfixes

• WindowsPath concatenation crash in encrypted_db.py (#99)
• deploy/cognithor.service for systemd (#CI fix)
• PREREQUISITES.md updated to cognithor
• Remaining python -m jarvis in tests fixed
• Graceful shutdown log event restored

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.87.1...v0.87.2

New Features

• cognithor config — Interactive CLI config TUI with dynamic model discovery (#95)
  • cognithor config list / get KEY / set KEY VALUE
  • Live model lists from Ollama, OpenAI, Gemini, LM Studio
  • Cached fallback registry (CI-updated weekly)
• contrib/ directory for community modules (Item 4 refactor)

Fixes

• Graceful CLI shutdown — no more CancelledError traceback on /quit (#96)
• 7 remaining from jarvis inline imports fixed (CI was failing)
• Ruff formatting for 17 files post-rename
• Version-pinned upgrade example in README (#94)
• Installer/bootstrap scripts updated for cognithor rename

Upgrade

pip install cognithor[all]==0.87.1

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.87.0...v0.87.1

BREAKING CHANGE: Package renamed from jarvis to cognithor

All imports change: from jarvis.X -> from cognithor.X

Migration

# Before
from jarvis.core.planner import Planner

# After
from cognithor.core.planner import Planner

Scope

• 1,265 files changed
• 4,474 Python import replacements
• 358 env var renames (JARVIS_ -> COGNITHOR_, both still work)
• 339 path renames (~/.jarvis -> ~/.cognithor, fallback to old path)
• 599 Flutter package renames (jarvis_ui -> cognithor_ui)

Backwards Compatibility

• jarvis CLI command still works (alias)
• Both COGNITHOR_* and JARVIS_* env vars accepted
• ~/.cognithor preferred, ~/.jarvis fallback
• "Jarvis" personality name unchanged

Also included (from v0.86.4)

• AST-based Python/Shell security analysis
• _safe_call subsystem failure tracking
• README accuracy overhaul
• Evolution Engine research optimization

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.86.4...v0.87.0

v0.86.4 — Security AST, _safe_call, README overhaul, Evolution fixes

Security (Item 5 — Reddit Refactor)

• Python AST Guard: Detects getattr/setattr bypass, chr() concat, __import__, exec/eval, dunder escape (40 tests)
• Shell AST Guard: bashlex parsing with regex fallback, 60+ blocked commands, substitution/chaining detection (48 tests)
• Gatekeeper: AST Layer 1 + Regex Layer 2 dual-defense for both Python and Shell

Observability (Item 3 — Reddit Refactor)

• _safe_call: All 22 subsystem inits in advanced.py tracked with failure registry
• Health endpoint: Shows degraded status + failed subsystem names
• 11 unit tests for safe_call module

README Accuracy (Item 2 — Reddit Refactor)

• Live CI + PyPI badges replace static ones
• Sandbox claims corrected (platform-adaptive, not "4-level")
• Tool count 120+, installation extras documented
• Development Status downgraded to Beta

Evolution Engine

• LLM-generated search queries instead of static templates
• search_and_read instead of web_search for source discovery
• Coverage thresholds increased, proportional scoring (1% granularity)
• Back-to-back cycles (5s pause instead of 5min cooldown)
• Journal API endpoint, Evolution config page in Flutter UI
• Learning goals persist across config saves

Reddit Lead Hunter

• Hard routing bypasses Planner for reddit_scan
• Product parameter added to reddit_scan
• LLM scoring uses qwen3:32b (was missing model parameter)

Kanban

• pending_review status for human-in-the-loop
• Scheduled tasks panel with cron job toggle
• Drag-and-drop reorder within columns

Fixes

• Bootstrap PEP 563 annotation resolution
• Config save: social/kanban/vault sections now editable
• ModelsConfig field names corrected
• DB corruption from force-kills handled
• Auto-upgrade syncs code + UI + procedures

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.86.3...v0.86.4

v0.86.3 — Config fixes, Evolution goals, Reddit scoring

Fixes

• Reddit scanner LLM call: Was missing model parameter — all 500 posts scored 0. Now uses qwen3:32b
• ModelsConfig crash: models.default.name doesn't exist, changed to models.planner.name
• Config save: social, kanban, vault sections were not in editable sections list — toggles didn't persist
• Evolution Goals API: Added POST (create single goal) and PATCH (update by ID) endpoints. GET now returns structured objects

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.86.2...v0.86.3

v0.86.2

Reddit Lead Hunter

• Hard routing: Gateway bypasses Planner, directly calls reddit_scan
• Product parameter added to reddit_scan
• JSON Schema fix for all 6 tools

Kanban Board

• New pending_review status (human-in-the-loop)
• Scheduled Tasks panel with pause/resume
• Drag-and-drop reorder within columns

Fixes

• Bootstrap endpoint PEP 563 fix
• Identity tools registered (4 new)
• Auto-upgrade syncs code + UI + procedures
• requirements.txt synced
• F821 active_skill scope fix

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.86.1...v0.86.2

What's Changed

Bugfixes

• Installer runtime error — Fixed Inno Setup {userprofile} constant (now {%USERPROFILE}) (#90)
• Logo not centered — Centered logo in collapsed sidebar rail (#91)
• Stale version banner — Updated hardcoded v0.80.0 to v0.86.x in start_cognithor.bat (#92)
• Reddit Skill unavailable — RedditLeadService now always initializes so MCP tools are registered even without product config
• Ruff formatting — Fixed lint errors in 4 files

New Features

• Kanban Scheduled Tasks — New "Scheduled" tab showing cron jobs with live next_run countdown, pause/resume toggle
• Drag-and-drop reorder — Reorder tasks within Kanban columns (sort_order persisted)
• Auto-upgrade from source — Installer now detects newer source trees and auto-upgrades on launch
• Bootstrap download progress — Verbose progress bar during Flutter Web UI download (#93)
• Morning Briefing expansion — Enhanced Day Planner procedure
• System architecture docs — Complete ASCII schematic of entire system
• CI workflow_dispatch — Flutter Web + Windows installer builds can now be triggered manually

API Changes

• — Pause/resume individual cron jobs
• — Jobs with next_run times
• — Batch sort_order updates

Full Changelog: https://github.com/Alex8791-cyber/cognithor/compare/v0.86.0...v0.86.1

Cognithor v0.86.0 -- Lead Wizard, Playwright Auto-Post, Deep Audit Fixes

Lead Wizard (Flutter v2)

• Full-screen sequential lead processor with progress bar
• Keyboard shortcuts: A=Archive, S=Skip, R=Reply, I=Improve (disabled inside TextField)
• Integrated RefinePanel: LLM improves draft with hint + 3 style variants
• TemplatePicker: select from saved reply templates
• Summary screen at completion: replied/skipped/archived counts
• "Process Queue" button on Leads tab (filters new leads, sorts by score)

Playwright Auto-Post

• Real browser automation with Reddit session cookie persistence
• Multiple selector strategies for Reddit UI changes
• Graceful clipboard fallback on any failure
• BrowserAgent wired in gateway post-init

Performance + Feedback in Detail Sheet

• PerformanceBadge: color-coded engagement score for replied leads
• FeedbackDialog: 5 tags (converted/conversation/ignored/negative/deleted)
• RefinePanel integrated into LeadDetailSheet

Deep Audit Fixes (2 Critical + 5 Important)

1. CRITICAL: ReplyRefiner LLM never wired -- refine was completely broken
2. CRITICAL: PerformanceTracker author_replied always False
3. Wizard keyboard shortcuts fired inside TextField
4. R=Reply shortcut documented but not implemented
5. Flutter scanInterval default 60 != backend 30
6. Skill file tools_required expanded to all 6 MCP tools

Pre-Release Verification

• Ruff lint: 0 errors
• Flutter analyze: 0 issues
• Python tests: 13,117 passed, 0 failed
• Flutter web build: successful

Cognithor v0.85.0 -- Intelligence Layer: Self-Improving Reddit Lead Hunter

LLM Reply Refinement

• refine() improves drafts with style profile + few-shot context + user hints
• generate_variants() creates 3 distinct reply styles (technical/casual/question)
• MCP tool: reddit_refine, REST: POST /leads/{id}/refine

Performance Tracking

• Automatic re-scanning every 6h: finds our comment, counts upvotes + replies
• Manual feedback tags: converted, conversation, deleted, ignored, negative
• Engagement score 0-100: upvotes3 + replies5 + author_replied10 + converted20

Subreddit Discovery

• LLM generates 20 candidates from product description
• Reddit JSON validation: subscribers, activity, sample posts
• Ranked by relevance x activity x log(subscribers), top 10 returned

Reply Templates

• Auto-save high-performing replies (engagement > 85) as templates
• Template variables: {product_name}, {subreddit}, {author}
• CRUD via MCP + REST

Feedback Learning Loop

• Weekly cron: LLM analyzes top-5 vs bottom-5 replies per subreddit
• Generates SubredditStyleProfile: what works, what fails, optimal tone/length
• Few-shot injection: top 3 replies as examples in every new draft
• Auto-adjusts reply_tone based on engagement trends

Scanner Intelligence

• REPLY_PROMPT now includes style context + few-shot examples
• Drafts improve with every learning cycle

New Infrastructure

• 5 new modules: refiner.py, tracker.py, discovery.py, templates.py, learner.py
• 3 new SQLite tables: reply_performance, reply_templates, subreddit_profiles
• 6 MCP tools total (3 new: refine, discover, templates)
• 14 REST endpoints total (8 new)
• 3 cron jobs (scan 30min, tracker 6h, learner weekly)

Stats

• 13,117 tests passing (34 new social tests)
• 0 ruff lint errors / 0 flutter analyze issues

Cognithor v0.84.0 -- Reddit Leads Tab, Deep Audit Fixes, Full Flutter Integration

Reddit Leads Flutter Tab (7th Screen)

• New "Leads" tab in main navigation (Ctrl+7)
• Stats bar: New / Reviewed / Replied counts
• Filter row: All / New / Reviewed / Replied (SegmentedButton)
• LeadCard widget: score badge, subreddit, status chip, meta row, action buttons
• LeadDetailSheet: full post, editable reply editor, Post Reply / Copy / Open on Reddit / Archive
• Scan Now FAB with loading spinner
• RedditLeadsProvider with 30s polling, filtering, CRUD
• 6 API client methods for leads endpoints
• 21 new i18n keys in EN/DE/ZH/AR

Social Listening Config Page

• Config > System > Social Listening
• Product name, description, subreddits, min score, scan interval, reply tone, auto-post
• Orange warning banner when product/subreddits not configured

Critical Bug Fixes (from Deep Audit)

3 showstopper bugs that made the Reddit Lead Hunter completely non-functional:

1. MCP tools never registered (Phase D ran before Phase F created the service)
2. LLM function always None (scanner could not score or draft)
3. Gatekeeper blocked all reddit tools at ORANGE risk

• 3 important fixes: default subreddits, min_score override, Flutter default alignment

Skill Chat Flow

• Skill now asks user for product + subreddits before scanning
• Suggests relevant subreddits based on product

Stats

• 13,083 tests passing
• 0 ruff lint errors / 0 flutter analyze issues
• 45-point verification script (scripts/verify_all.py)

Reddit Lead Hunter full integration: 3 trigger paths (Chat/Cron/REST), 7 new modules in social/, 6 REST endpoints, 34 new tests. No Reddit API key needed. Flutter Social Config page (8 fields, 11 i18n keys). Bootstrap auto-rebuilds Flutter UI when sources change. 13,083 tests passing.

Cognithor v0.82.0 — Robot Office Live, Deep Learning Pipeline, 13 Bug Fixes

Robot Office Live Wiring (#84)

• Dynamic Robots — 3 PGE Trinity (Planner, Executor, Gatekeeper) always present + dynamic user agents from configured agents list
• Real-time PGE State Sync — Planner types when planning, Executor works when running tools, Gatekeeper reviews — all via WebSocket events
• System Metrics — Server rack LEDs pulse with real CPU usage, ceiling lights dim with system load
• Kanban Board — Real colored dots per status column (grey/blue/green/red) with hover tooltips showing actual task titles
• System Glow — PGE Trinity robots have subtle pulsing glow to distinguish from user agents
• Real Task Messages — "Planning: [user message]", "Running: [tool_name]", "Reviewing: [tool]" instead of fake messages

Deep Learning Upload Pipeline (#89)

• Hybrid Processing — Immediate chunk-indexing + background KnowledgeBuilder pipeline (Vault, entities, identity memory, LLM summary)
• Priority Queue — High/Normal/Low priority with in-memory heapq ordering
• PDF Vision — Image-heavy PDF pages analyzed via vision model
• OCR Fallback — Tesseract OCR for scanned PDFs when text extraction yields < 100 chars
• YouTube Frames — Key frame extraction via yt-dlp + ffmpeg for HIGH priority uploads
• Evolution Integration — Uploaded material seeds autonomous learning during idle cycles
• Flutter UI — Priority dropdown on Teach screen, deep-learn queue panel, progress events

Bug Fixes

• #78 — Windows uninstaller script (code-only or full removal with DELETE confirmation)
• #79 — Removed duplicate Evolution settings page (1199 lines)
• #80 — Translate Prompts via Ollama now actually sends prompts to backend
• #81/#82 — Missing i18n in admin pages, sidebar (7 new keys in 4 languages)
• #83 — Logo fallback improved from plain "C" to brain icon with gradient
• #85 — Version mismatch (init.py stuck at 0.71.0) — already fixed in v0.81.0
• #86 — Operation Mode now has description explaining what each mode does
• #87 — QR pairing screen fetches payload from server instead of asking user
• #88 — Robot Office task messages replaced with clearly playful/decorative text

Stats

• 13,042 tests passing (Python 3.12/3.13 x Ubuntu/Windows)
• 0 ruff lint errors / 0 flutter analyze issues
• Flutter web build included (fresh)
• Closes: #78, #79, #80, #81, #82, #83, #84, #85, #86, #87, #88, #89

Cognithor v0.81.0 — UI Polish, i18n, OpenRouter, Lint Zero

New Features

• Token/Model Info on Chat Messages — each assistant response shows model name, backend, input/output token count, and duration (#70)
• OpenRouter Backend Support — full integration in backend switch dialog, API status check, and routing (#63)
• Agent Delete Button — Administration > Agents now has delete with confirmation (#69)
• Incognito Exit — clickable badge in AppBar + toggle in drawer to leave incognito mode (#67)
• Active Model Display — Provider status card shows current planner/executor model names (#74)
• Installer Downgrade Protection — bootstrap and Inno Setup detect and warn on version downgrade

Bug Fixes

• 299 Ruff Lint Errors Resolved — E501, F841, B904, SIM102, RUF006, and 15 more rule categories. Zero lint errors across all rules
• Flutter Analyze Clean — all 19 issues resolved (unused code, deprecated APIs, null-safety)
• Fail-Fast Backend Routing — non-Ollama backends no longer silently fall back to Ollama on creation failure. Clear error messages instead of confusing "HTTP None" (#74)
• Audit Verify Crash — config_manager was not passed to monitoring routes (#66)
• Search Navigation — Ctrl+K search now navigates to the correct config category + sub-tab (#72)
• Save Error Detail — config save errors show the actual message per section, not just section names (#71)
• Number Field Bounds — critical fields (token budget, timeouts) now have max limits (#71)
• API Key Display — saved keys show 24 bullet characters instead of "***", no useless eye button (#64)

UI Cleanup

• Removed Duplicate Pages — Models (#62) and Agents (#65) config pages removed; Administration pages are the single source
• Removed Legacy React/Preact UI — Flutter Migration Phase 4 complete. ui/ and apps/pwa/ deleted (~2,500 LOC removed)
• Unified Backend Dialog — all 17 providers shown (was 5 hardcoded), valid list derived from config Literal type (#63)
• Kanban Dynamic Agents — agent dropdown shows only configured agents, not hardcoded list (#73)
• Version Consistency — init.py, cognithor.iss, first_run.py all synced; first_run reads version dynamically

Internationalization

• Vault Config Page — 10 new keys in EN/DE/ZH/AR (#68)
• Kanban UI — 22 new keys (column names, config dialog, task sources) (#75)
• Device Pairing — 35 missing ZH/AR translations added
• 0 Untranslated Keys across all 4 languages (EN/DE/ZH/AR)

Stats

• 13,006 tests passing (Python 3.12/3.13 x Ubuntu/Windows)
• 0 ruff lint errors
• 0 flutter analyze issues
• ~199k LOC source / ~162k LOC tests / ~55k LOC Flutter
• Closes: #62, #63, #64, #65, #66, #67, #68, #69, #70, #71, #72, #73, #74, #75

Fixes

• #60: Install scripts now ask before installing Ollama and pulling models
• #61: start_cognithor.bat keeps window open on error, removed problematic ASCII-art
• CI: Fixed test importing renamed _TOOL_STATUS_KEYS

What's New in v0.80.0

Phase 3 HybridClaw (8 Features)

• Tool-Loop-Detection — SHA-256 based endless loop prevention
• 3-Tier Context Guard — in-flight context window protection
• Security Extension Hooks — exfiltration detection
• Config-Versioning — snapshot + rollback
• Stalled-Turn-Counter — MAX=20 safety cap
• In-Loop Compaction — LLM + heuristic context trimming
• Model-Retry + Stream-Fallback — graceful LLM error recovery
• Concierge Routing — urgency-based request routing

Ralph Agent-Loop

• CONTINUE/STOP prompt pattern for multi-step autonomous tasks
• Progress detection, budget limits, iteration tracking

Adaptive Skill Performance Tracking

• Sliding window degradation detection
• Auto-disable + cooldown re-enable
• REST endpoints for skill health monitoring

Skill Lifecycle

• Generated skills (~/.jarvis/skills/generated/) now loaded into registry

Evolution Phase 5 — Autonomous Deep Learning

• AutonomousLearner — curiosity-driven knowledge gap identification
• EvolutionRAG — lightweight SQLite-backed RAG pipeline
• EvolutionScheduler — cron-based recurring learning tasks
• MetaLearner — strategy analysis + self-improvement

i18n Phase 2

• 36 keys migrated across gateway, planner, gatekeeper (en/de/zh)
• SHA-256 integrity verification for all locale packs

Flutter UI

• Connected Devices — list, pair, revoke mobile devices + QR payload
• Network Settings — interface detection, per-interface toggle, auto-detect
• QR Scanner — camera placeholder + paste fallback for pairing

Mobile Pairing & Networking

• Smart host binding (auto-detect Tailscale/ZeroTier/WireGuard)
• Multi-endpoint network settings with persistent config
• Device pairing with HMAC tokens (180d TTL)

Tests

• ~270 new tests across all features

What's New

Phase 3: HybridClaw Intelligence (8 features)

• Tool-Loop-Detection: SHA-256 based, detects generic_repeat + ping_pong patterns
• 3-Tier Context Guard: per-result truncation, old result compaction, overflow flag
• Security Extension Hooks: exfiltration detection, binary office write blocking
• Config Versioning: automatic snapshots + rollback via REST API
• Stalled-Turn Counter: MAX=20 turns without progress, auto-recovery
• In-Loop Compaction: LLM-based mid-session summarization with heuristic fallback
• Model-Retry + Stream Fallback: exponential backoff, auto stream-to-sync on errors
• Concierge Routing: urgency-based model selection (asap/balanced/no_hurry)

Mobile Pairing

• QR-code based device pairing with HMAC tokens (180-day expiry)
• REST API: /devices, /devices/pair, /devices/{id} (revoke)
• Connected Devices management

Tailscale Auto-Detection

• Automatic Tailscale interface detection (100.64.0.0/10 CGNAT)
• Smart API binding: localhost when local, 0.0.0.0 when Tailscale present
• Mobile apps connect from anywhere via Tailscale tunnel
• No manual --api-host configuration needed

Security (from v0.78.2)

• GHSA-cognithor-001 fix: bootstrap token via HTML meta tag
• All API endpoints require Bearer auth
• Default bind 127.0.0.1 (unless Tailscale detected)

~230 new tests this release cycle

Security Fix

GHSA-cognithor-001 — Unauthenticated Master Token Disclosure (CVSS 9.8 CRITICAL)

The /api/v1/bootstrap endpoint returned the master bearer token without authentication.

Fixes

• Bootstrap endpoint restricted to loopback addresses only
• Default API bind changed from 0.0.0.0 to 127.0.0.1

All users should upgrade immediately.

Reported by Offgrid Security

SUPERSEDED — Upgrade to v0.78.2

This release contains a critical security vulnerability (GHSA-cognithor-001). Please use v0.78.2 instead.

What's New

Bug Fixes

• Launcher health-check polling -- browser now waits for backend to be ready before opening (#46)
• Encrypted file init -- honors JARVIS_DB_KEY env var even without config.yaml
• sqlite3.Row compatibility -- fixes cross-platform persistence issue
• Idle detector threshold -- fixes race condition with threshold=0
• RAM detection test -- graceful fallback when psutil not installed
• Flutter Kanban provider -- fixed ApiClient integration (was using raw HTTP instead of project client)
• Setup wizard -- fixed String? type mismatch
• Installer quoting -- fixed Inno Setup path handling for directories with spaces

Improvements

• Duplicate instance prevention -- checks port 8741 before starting
• Silent startup -- clean professional output, no library warnings
• Generic LLM banner -- not Ollama-specific
• Post-install UX -- shows startup progress instead of running hidden

CI

• All 5 jobs green (Ruff lint + Python 3.12/3.13 x Ubuntu/Windows)
• ~12,600 tests passing

Downloads

• Windows Installer: CognithorSetup-0.78.0.exe (see Assets below)
• Android APK: Built via GitHub Actions (build-mobile workflow)
• iOS IPA: Built via GitHub Actions (build-mobile workflow)

Full Changelog: v0.77.1...v0.78.0

Changes since v0.77.0

Code Audit (3 bugs fixed)

• kanban_tools: None dereference guard in _handle_update
• engine: transition() guaranteed non-None return
• deep_learner: run_quality_test type mismatch fixed (SubGoal, not LearningPlan)

Flutter UI

• Evolution Goals/Plans/Journal page in Admin Hub
• Kanban Config Dialog (task sources, guards, defaults)
• Generated i18n files updated with 18 missing keys

System Check

• 151/151 tests passing
• Ruff lint: 0 errors
• skill_registry wiring bug fixed (F821)
• ISS compiler type mismatch + path quoting fixed

Downloads

• CognithorSetup-0.77.0.exe — Full installer with all v0.77.x fixes

What's New in v0.77.0

Evolution Engine Phase 5 Hardening

• Autonomous Exam-Based Learning Cycle: CycleController manages Expand -> Examine -> Decide loop
  • Every 10 horizon expansions: automatic quality exam (LLM-generated questions, self-answered)
  • Score >= 0.8: goal marked as MASTERED, learning stops
  • Score stagnating (delta < 0.05 over 2 exams): frequency reduced to 25%, Kanban task created
  • Score recovers: automatically returns to full frequency
• ATL Action Dispatch Complete: goal_management (add/pause/resume/complete) + file_management (reports, notes)
• Risk Ceiling Enforcement: GREEN/YELLOW risk tiers enforced before action dispatch
• HorizonScanner Targeted Expansion: exam gaps feed directly into next expansion cycle
• Evolution REST API: 8 endpoints for goals, plans, journal, stats
• 20 new tests

Computer Use Phase 2 — Vision-Guided Interaction

• computer_click_element: Click UI elements by description instead of coordinates
• computer_wait_for_change: Detect if screen changed after an action
• 9 new tests

Complete Windows Uninstaller

• Registered in Windows Apps & Features (Settings > Apps)
• Stops Ollama before removal
• Asks user: keep or delete user data (~/.jarvis/)
• Cleans PATH, shortcuts, registry

i18n Migration Complete

• 314 hardcoded German strings migrated to t() calls
• Flutter screen strings migrated to AppLocalizations

Also in this release

• Skill Lifecycle fix: generated skills auto-loaded + hot-reload
• Kanban auto-task creation wired into Cron, Evolution, Executor

Changes since v0.76.0

i18n Migration Phase 2 (Complete)

• 314 hardcoded German strings migrated to t() calls across 20 MCP/Core files
• Zero hardcoded user-facing German strings remaining in MCP + Core
• All strings available in EN/DE/ZH
• SHA-256 integrity hashes regenerated
• 20 missing Flutter ARB keys added to de/zh/ar

Skill Lifecycle Fix

• Generated skills now always loaded into registry (was silently skipped)
• Hot-reload: skills created during runtime immediately available
• SkillGenerator wired to SkillRegistry for auto-registration

Kanban Integration

• Auto-task creation from Cron jobs, Evolution Engine, and Executor wired
• Tasks automatically created on repeated tool failures (via RecoveryEngine)

What's New in v0.76.0

Interactive Kanban Board

• New 6th Tab in Flutter Command Center with drag-and-drop task management
• Dual-View: Toggle between "My Tasks" board and "Live Pipeline" monitor
• 6 Task Sources: Manual UI, Chat detection, Cron jobs, Evolution Engine, Agent-created, System-generated
• Sub-Task Support: Parent-child relationships with cascade cancel and auto-verification
• Status Machine: TODO -> IN_PROGRESS -> VERIFYING -> DONE with enforced transitions
• 3 MCP Tools: Agents can create, update, and list their own tasks
• 10 REST Endpoints: Full CRUD + move + history + stats + config
• SQLCipher Encrypted Storage: All tasks encrypted at rest
• Guards: Max auto-tasks per session, max subtask depth, WS broadcast debounce
• Configurable via Flutter UI: Toggle sources, set limits, manage columns and labels
• 63 Backend Tests: Models, Store, Engine, Sources, API, MCP Tools

Claw Code-Inspired Improvements

• V5 Context-Window Preflight: Checks token count before every LLM call, auto-compacts if exceeded
• V3 Failure-Taxonomy: 8 failure classes with structured recovery recipes and escalation policies
• V6 Per-Tool Permissions: Tools declare risk level at registration, Gatekeeper reads annotations

i18n Fixes

• Regenerated stale SHA-256 hash files (integrity checks were failing)
• Fixed de.json version mismatch (1.0.0 -> v0.35.5)
• Added 20 missing ARB keys to de/zh/ar Flutter locales

Previous (v0.75.x)

• Windows Installer with First-Run Setup Wizard
• Database encryption by default (SQLCipher + OS Keyring)
• Clean professional startup output (zero warnings)
• Auto-generated encryption keys
• Fixed ddgs import, credential/sqlcipher warning spam

Changes since v0.75.0

Clean Professional Startup

• Zero warnings, zero debug noise on startup
• Clean output: Banner + [OK] lines + "Cognithor is ready."
• All third-party library noise suppressed (FAISS, HF Hub, tqdm, SQLCipher)

Database Encryption by Default

• SQLCipher encryption enabled for all new installations
• Switched from pysqlcipher3 (needs compilation) to sqlcipher3 (pre-built wheels)
• Auto-generated encryption key stored in OS keyring
• Credential store auto-generates passphrase via keyring

Claw Code-Inspired Improvements

• V5 Context-Window Preflight: Checks token count before every LLM call, auto-compacts if exceeded
• V3 Failure-Taxonomy: 8 failure classes with structured recovery recipes
• V6 Per-Tool Permissions: Tools declare risk level at registration, Gatekeeper reads annotations

Bug Fixes

• Fix ddgs import name (duckduckgo_search -> ddgs)
• Fix credential/sqlcipher warning spam (1x instead of 20x)
• Fix encrypted procedures summary (1 line instead of 60)
• Fix iOS IPA build bundle ID verification

Downloads

• Windows: CognithorSetup-0.75.1.exe below
• iOS: IPA will be attached when CI build completes

Windows Installer & First-Run Wizard

Windows Installer (One-Click Setup)

• Download CognithorSetup-0.75.0.exe below
• Bundles embedded Python 3.12, Ollama, and Flutter Command Center
• No admin privileges required

First-Run Setup Wizard

• Hardware auto-detection (CPU, RAM, GPU/VRAM)
• LLM provider choice: Ollama / External API / Both
• Model recommendation based on hardware
• Language selection (EN/DE/ZH)
• 19 community skills + 6 default agents

Bug Fixes

• Fix ddgs import check (v9.x rename)
• Fix credential/sqlcipher warning spam (1x instead of 20x)
• Fix encrypted procedures spam (summary instead of per-file)
• Fix Flutter UI path detection for installed mode