# Features

The Dynamic Skill Creator helps you build a personalized knowledge base for Claude Code through these core capabilities:

| Capability | Description |
|------------|-------------|
| **1. Capturing Patterns** | Observes your Claude Code sessions to detect recurring workflows, commands, and file access patterns |
| **2. Suggesting Skills** | Proposes skill creation when patterns repeat 3+ times with evidence explaining why |
| **3. Managing Skills** | Provides guided workflows for creating, searching, listing, and organizing skills |
| **4. Auto-Loading** | Automatically loads relevant skills based on context while respecting token budgets (2-5% of context) |
| **5. Learning** | Refines skills based on your corrections and feedback with bounded parameters and user confirmation |
| **6. Composing Agents** | Groups frequently co-activated skills into composite agents stored in `.claude/agents/` |
| **7. Quality Validation** | Detects semantic conflicts between skills and scores activation likelihood (v1.1) |
| **8. Testing & Simulation** | Automated test cases, activation simulation, and calibration benchmarks (v1.2) |
| **9. Agent Teams** | Multi-agent team coordination with leader-worker, pipeline, and swarm topologies (v1.4) |
| **10. Pattern Discovery** | Scan session logs to discover recurring workflows and generate draft skills automatically (v1.5) |
| **11. Orchestrator** | Master agent routing user intent to GSD commands via dynamic discovery and intent classification (v1.7) |
| **12. Skill Workflows** | Multi-step skill chains with dependency tracking and crash recovery (v1.7) |
| **13. Skill Roles** | Behavioral constraints and tool scoping for agent personas (v1.7) |
| **14. Work Bundles** | Project-phase skill sets with progress tracking and auto-suggestion (v1.7) |
| **15. Inter-Skill Events** | Event emit/listen system enabling causal activation chains (v1.7) |
| **16. Skill Pipeline** | Composable pipeline architecture with pluggable stages replacing monolithic skill loading (v1.8) |
| **17. Token Budgets** | Per-agent token budget profiles with critical/standard/optional priority tiers (v1.8) |
| **18. Capability Planning** | Auto-generated capability manifests, phase declarations, and skill injection into executors (v1.8) |
| **19. Cache Optimization** | Cache-aware skill ordering with cacheTier metadata for prompt cache efficiency (v1.8) |
| **20. Research Compression** | 10-20x research document reduction with staleness detection (v1.8) |
| **21. Parallelization Advisor** | Wave-based parallel execution recommendations from plan dependency analysis (v1.8) |
| **22. Spec Alignment** | Full Claude Code spec compliance: $ARGUMENTS injection, context:fork, dual-format allowed-tools, shell injection prevention (v1.9) |
| **23. Progressive Disclosure** | Large skills auto-decompose into SKILL.md + references/ + scripts/ with token budget awareness (v1.9) |
| **24. Cross-Platform Portability** | Export skills as portable archives or platform-specific formats (Claude, Cursor, Codex, Copilot, Gemini) (v1.9) |
| **25. Evaluator-Optimizer** | Precision/recall/F1 tracking, A/B evaluation with t-test significance, health dashboard (v1.9) |
| **26. MCP Distribution** | Publish .tar.gz skill packages, install from local/remote, MCP server with 4 tools (v1.9) |
| **27. Enhanced Topologies** | Router and Map-Reduce team patterns, inter-team deadlock detection, cost estimation (v1.9) |
| **28. Session Continuity** | Save/restore/handoff with warm-start context and cross-session ephemeral promotion (v1.9) |
| **29. Agentic RAG** | Adaptive TF-IDF/embedding routing, corrective refinement, cross-project skill discovery (v1.9) |
| **30. Quality of Life** | Description quality scoring, budget dashboard, Mermaid dependency graphs, GSD command injection (v1.9) |
| **31. Security Hardening** | Path traversal prevention, YAML safe deserialization, JSONL integrity with checksums, secret redaction, dangerous command deny list (v1.10) |
| **32. Data Integrity** | SHA-256 checksums on JSONL entries, schema validation, observation rate limiting, anomaly detection, compaction and purge (v1.10) |
| **33. Learning Safety** | Cumulative drift tracking with 60% threshold, contradictory feedback detection, skill audit CLI (v1.10) |
| **34. Access Control** | File integrity monitoring, audit logging, inheritance depth limits, impact analysis, concurrency locks, operation cooldowns (v1.10) |
| **35. Operational Safety** | Hook error boundaries, hook safety validation, orchestrator confirmation gates, classification audit logging (v1.10) |
| **36. GSD Integration Config** | Per-feature toggles, Zod-validated JSON config, CLI validation command, opt-out model with sensible defaults (v1.11) |
| **37. Install & Git Hooks** | Idempotent install script with --uninstall, POSIX shell post-commit hook for zero-cost commit observation to sessions.jsonl (v1.11) |
| **38. Session Commands** | `/sc:start` warm-start briefing, `/sc:status` budget dashboard, `/sc:suggest` interactive review, `/sc:observe` session snapshot, `/sc:digest` learning digest (v1.11) |
| **39. Wrapper Commands** | `/wrap:execute`, `/wrap:verify`, `/wrap:plan` with skill loading, `/wrap:phase` smart lifecycle router (v1.11) |
| **40. Passive Monitoring** | Plan-vs-summary diffing, STATE.md transition detection, ROADMAP.md structural diff, scan-on-demand architecture (v1.11) |
| **41. Planning Docs Dashboard** | Markdown-to-HTML generator parsing `.planning/` artifacts into 5 browsable pages with dark theme, no external dependencies (v1.12) |
| **42. Structured Data & SEO** | JSON-LD (Schema.org), Open Graph meta tags, semantic HTML5 on all dashboard pages (v1.12) |
| **43. Incremental Builds** | SHA-256 content hashing, build manifest, auto-refresh with scroll preservation, visual refresh indicator (v1.12) |
| **44. Live Session Pulse** | Real-time session monitoring with ticking duration, commit feed, heartbeat indicator, message counters at 1-2s sample rate (v1.12.1) |
| **45. Phase Velocity & Planning Quality** | Timeline visualization, per-phase stats, TDD rhythm analysis, accuracy scores, emergent work ratio at 5-10s sample rate (v1.12.1) |
| **46. Historical Trends** | Milestone comparison, commit type distribution, velocity curves, file hotspots with CSS-only visualizations (v1.12.1) |
| **47. Message Stack** | Async command queuing with priority levels (push/pop/poke/drain) and tmux session integration (v1.13) |
| **48. Session Lifecycle** | Managed Claude Code sessions via tmux with start/list/watch/pause/resume/stop/save and heartbeat monitoring (v1.13) |
| **49. Recording & Playback** | Stream capture (terminal, stack events, file changes) with 4 replay modes (analyze, step, run, feed) and 14-metric computation (v1.13) |
| **50. Pipeline List Coprocessor** | Declarative workflow programs with WAIT/MOVE/SKIP instructions synchronized to GSD lifecycle events (v1.13) |
| **51. Offload Engine** | Script promotion from skill metadata for deterministic operations executed outside the context window (v1.13) |
| **52. Team-as-Chip Framework** | Amiga-inspired architecture with 4 specialized chips (Agnus, Denise, Paula, Gary), FIFO message ports, 32-bit signals (v1.13) |
| **53. Exec Kernel** | Prioritized round-robin scheduler with 18 typed message protocols and per-team token budgets with burst mode (v1.13) |
| **54. Pipeline Learning** | Observation-to-list compiler, Jaccard feedback engine, versioned library with best-match retrieval (v1.13) |
| **55. Execution Capture** | Pipeline pairing tool_use/tool_result with SHA-256 content hashes for cross-session comparison (v1.14) |
| **56. Determinism Analysis** | Three-tier classification (deterministic/semi/non) with configurable variance thresholds and cross-session stability (v1.14) |
| **57. Promotion Pipeline** | Weighted composite scoring, script generation with dry-run validation, F1/MCC gatekeeper, drift monitoring with auto-demotion (v1.14) |
| **58. Lineage Tracking** | Bidirectional provenance querying from observation through pattern to promotion to script across all pipeline stages (v1.14) |
| **59. Terminal Integration** | Wetty browser-based terminal with tmux session binding, health checks, dark theme, and config-driven URLs (v1.15) |
| **60. Unified Dev Launcher** | DevEnvironmentManager composing dashboard + terminal with single start/stop/status API (v1.15) |
| **61. Filesystem Message Bus** | Bidirectional `.planning/console/` with Zod-validated JSON envelopes, inbox/outbox routing, and pending→acknowledged lifecycle (v1.16) |
| **62. Upload & Configuration** | Drag-and-drop markdown ingestion, document metadata extraction, and 7-section milestone configuration form (v1.16) |
| **63. Interactive Questions** | 5 question types (binary, choice, multi-select, text, confirmation) with timeout fallback and urgency escalation (v1.16) |
| **64. Console Dashboard** | Live session status, hot-configurable settings, activity timeline, and clipboard fallback mode (v1.16) |
| **65. Staging Pipeline** | 5-state filesystem pipeline (inbox/checking/attention/ready/aside) with structured metadata and trust-aware reporting (v1.17) |
| **66. Hygiene Engine** | 11 built-in patterns detecting embedded instructions, hidden content, YAML config safety, with familiarity tiers and trust decay (v1.17) |
| **67. Smart Intake** | Clarity assessment routing (clear/gaps/confused), resource analysis with skill cross-reference, and derived knowledge checking (v1.17) |
| **68. Staging Queue** | 7-state machine with append-only audit log, cross-queue dependencies, pre-wiring engine, and dashboard queue panel (v1.17) |
| **69. CSS Design System** | 6 domain colors, 4 signal colors, typography tokens (Inter + JetBrains Mono), spacing scale, and 5 status states (v1.18) |
| **70. Gantry & Topology** | Persistent status strip with agent circles/phase fractions/budget bar, and subway-map topology with SVG bezier edges (v1.18) |
| **71. Entity Shapes & Activity** | 6 SVG entity shapes with dual encoding, collapsible legend, activity feed with Unicode indicators and domain colors (v1.18) |
| **72. Domain Identifiers** | Domain-prefixed encoding (F-1, B-1.api, T-1:rcp) with backward compatibility and SKILL.md metadata persistence (v1.18) |
| **73. Budget Inventory Model** | LoadingProjection separating installed total from loadable total with tier-aware selection and profile awareness (v1.19) |
| **74. CLI Status Redesign** | Two-section layout with proportional sizing, loading projection, color-coded budget bar, and JSON output mode (v1.19) |
| **75. Budget Configuration** | Per-profile cumulative budgets in integration config, env var backward compat, dual-dimension history with migration (v1.19) |
| **76. Unified CSS Pipeline** | 18 component style modules wired into dashboard generator with design system token compliance, no hardcoded hex (v1.20) |
| **77. Dashboard Data Collectors** | Topology, activity feed, budget-silicon, and staging queue collectors bridging real data to dashboard panels (v1.20) |
| **78. Console Page Assembly** | Console as 6th dashboard page with settings panel, activity timeline, question cards, and submit flow (v1.20) |
| **79. Tauri Desktop Shell** | Native app with Rust backend, Vite webview, bidirectional IPC (commands, events, channels), and capability ACL security (v1.21) |
| **80. WebGL CRT Engine** | Multi-pass post-processing with scanlines, barrel distortion, phosphor glow, chromatic aberration, vignette, and CSS fallback (v1.21) |
| **81. Indexed Palette System** | 32-color system with 5 retro-computing presets, OKLCH generation, and copper list raster effects (v1.21) |
| **82. Native PTY Terminal** | Rust-backed pseudo-terminal with xterm.js, watermark-based flow control, and tmux session binding (v1.21) |
| **83. Desktop Environment** | Amiga Workbench-inspired window manager with depth cycling, drag/resize, taskbar, pixel-art icons, system menu (v1.21) |
| **84. Boot & Calibration** | Three-screen calibration wizard, Amiga chipset boot sequence animation, accessibility auto-detection (v1.21) |
| **85. Local Cloud Infrastructure** | PXE boot server, kickstart automation, hypervisor-agnostic VM provisioning across KVM/VMware/VirtualBox (v1.22) |
| **86. Minecraft Knowledge World** | Fabric server with Syncmatica mod stack, themed districts, spawn tutorial, schematic library, educational curriculum (v1.22) |
| **87. Platform Portability** | Hardware discovery, distribution abstraction (dnf/apt/pacman), multi-hypervisor VM operations, container fallback (v1.22) |
| **88. Amiga Emulation** | FS-UAE with AROS ROM, application profiles (Deluxe Paint, OctaMED, ProTracker, PPaint), format converters (v1.22) |
| **89. Chipset Formalization** | 20 SKILL.md files, 10 agents, 5 teams, 4 topologies, unified chipset YAML with trigger routing (v1.22) |
| **90. Operational Maturity** | RCON-quiesced backups with 24/7/4 rotation, Prometheus monitoring, golden image lifecycle, 4 runbooks (v1.22) |
| **91. AMIGA Mission Infrastructure** | MC-1 Control Surface, ME-1 Mission Environment, CE-1 Commons Engine, GL-1 Governance Layer with 4 typed ICDs (v1.23) |
| **92. Apollo AGC Simulator** | Complete Block II CPU emulation: 38 instructions, ones' complement ALU, bank-switched memory, interrupts, I/O channels (v1.23) |
| **93. AGC Executive/Waitlist** | Priority-based cooperative scheduler with 8 core sets, timer-driven task queue, BAILOUT restart protection (v1.23) |
| **94. DSKY Interface** | Authentic display model with relay decoding, 19-key keyboard, VERB/NOUN processor, Executive Monitor, learn mode (v1.23) |
| **95. AGC Development Tools** | yaYUL assembler, step debugger with breakpoints/watchpoints, disassembler, rope loader, 54-test validation suite (v1.23) |
| **96. AGC Curriculum** | 11 chapters from orientation to AGC-to-GSD patterns, 8 hands-on exercises, 1202 alarm capstone (v1.23) |
| **97. RFC Reference Skill** | 3-agent system (retriever, analyzer, citation builder), 5 Python scripts, 57-RFC index, multi-format output (v1.23) |
| **98. Observation Hook System** | Zod-validated learner observation types (activity completion, assessment results, time spent, pack lifecycle), ObservationEmitter with sink registration and error isolation (v1.27) |
| **99. AMIGA Event Bridge** | KnowledgeEventBridge converting LearnerObservation to AMIGA EventEnvelope with 6 event types (KNOWLEDGE_*), priority escalation for pack_complete, correlation ID tracking (v1.27) |
| **100. Learning Pattern Detector** | 4-pattern detection engine (sequence, timing, scoring, engagement) with confidence scoring, threshold filtering, and skill suggestion generation for high-confidence patterns (v1.27) |
| **101. Pathway Adapter** | Personalized learning path adaptation based on learner history: struggle detection (reinforcement), excellence detection (acceleration), traceable adaptation decisions with observation evidence (v1.27) |
| **102. Activity Scaffolder** | Pattern-to-activity generation creating new PackActivity objects from learning patterns, insertable into existing activity chains, pattern-specific generation for 4 pattern types (v1.27) |
| **103. Approach Promoter** | Pattern-to-skill promotion converting high-confidence learning patterns to reusable skill definitions with trigger conditions, action steps, and SKILL.md markdown output (v1.27) |
| **104. Knowledge Pack Integration** | Complete skill-creator integration infrastructure: observation hooks, AMIGA bridge, pattern detection, pathway adaptation, activity scaffolding, approach promotion, barrel exports with 144 new tests (v1.27) |
| **105. Filesystem Message Bus** | 8-priority ISA-encoded message bus with Dispatcher routing, queue health metrics, dead-letter handling, and message pruning (v1.28) |
| **106. 10-Position Staff Topology** | Coordinator, Relay, Planner, Configurator, Monitor, Executor, Verifier, Sentinel, Chronicler, Dashboard organized into 5 operational divisions (v1.28) |
| **107. Topology Profiles** | 4 auto-selected profiles (Scout/Patrol/Squadron/Fleet) based on phase count, with role subset hierarchy (v1.28) |
| **108. HALT/CLEAR Protocol** | Priority 0 emergency stop via bus message, 9-type recovery decision matrix in Sentinel position (v1.28) |
| **109. MNA Circuit Simulator** | DC/AC/transient analysis with Newton-Raphson solver, 7 component models, stamp logging for educational visibility (v1.29) |
| **110. Digital Logic Simulator** | 8 gate types with CMOS internals, truth tables, ASCII timing diagrams, flip-flops (SR/D/JK/T), ripple-carry adder (v1.29) |
| **111. Safety Warden** | 3 operating modes (annotate/gate/redirect), IEC 60449 voltage classification, positive framing with 8 prohibited words (v1.29) |
| **112. Electronics Learn Mode** | 3-level depth (practical/reference/mathematical), H&H citation lookup, depth markers in all 15 modules (v1.29) |
| **113. 77 Interactive Labs** | Hands-on electronics experiments across 15 modules in 4 tiers grounded in *The Art of Electronics* (v1.29) |
| **114. 5 Specialized Engines** | DSP (FFT/FIR), GPIO (UART/SPI/I2C), PLC (ladder logic/PID), Solar (MPPT/battery), PCB (impedance/DRC/Gerber) (v1.29) |
| **115. Vision-to-Mission Types** | Zod schemas for all 8 VTM document structures with inferred TypeScript types and 60/40 budget constraints (v1.30) |
| **116. Vision Document Parser** | Regex section extraction, archetype classifier (Educational/Infrastructure/Organizational/Creative), quality checker (v1.30) |
| **117. Wave Planner** | Parallel track detection via graph coloring, dependency graphs with critical path, sequential savings calculator (v1.30) |
| **118. Model Assignment Engine** | Weighted signal scoring for Opus/Sonnet/Haiku, budget validator, downgrade-only auto-rebalance (v1.30) |
| **119. Pipeline Orchestrator** | End-to-end vision → research → mission with stage skipping, template rendering, enrichment, structured errors (v1.30) |
| **120. MCP Foundation Types** | Shared TypeScript + Rust types for MCP structures with Zod schemas and serde round-trip parity (v1.31) |
| **121. Rust MCP Host Manager** | Server lifecycle with stdio transport, crash detection, exponential backoff, tool routing, capability caching, trace emission (v1.31) |
| **122. Gateway MCP Server** | 19 tools across 6 groups, 4 resource providers, 3 prompt templates, Streamable HTTP with bearer auth and scope enforcement (v1.31) |
| **123. MCP Template Generators** | Server, host, and client project scaffolds generating complete, buildable TypeScript projects (v1.31) |
| **124. Agent Bridge** | SCOUT and VERIFY as MCP servers, EXEC as MCP client, inter-agent communication with staging gates and context isolation (v1.31) |
| **125. MCP Security Pipeline** | Hash verification, trust lifecycle with 30-day decay, invocation validation, rate limiting, audit logging, staging gates on all paths (v1.31) |
| **126. Blueprint Editor MCP Blocks** | Server/Tool/Resource block types with port rendering, type-safe wiring engine, deny-by-default rules (v1.31) |
| **127. MCP Trace & Security Panels** | Real-time JSON-RPC trace with sparklines, Security Dashboard with trust states and blocked call log (v1.31) |
| **128. Brainstorm Rules Engine** | Osborn's 4 rules enforced architecturally with Critic gate, two-stage evaluative detection (<5% FP), Black Hat constraint, violation logging (v1.32) |
| **129. Session Manager & Phase Controller** | 5-status state machine, strict 5-phase ordering (Explore→Diverge→Organize→Converge→Act), agent activation matrix, timer system (v1.32) |
| **130. 16 Brainstorming Techniques** | Individual (freewriting, mind mapping, rapid ideation, question brainstorming), collaborative (brainwriting 6-3-5, round robin, brain-netting, rolestorming, figure storming), analytical (SCAMPER, Six Hats, starbursting, Five Whys), visual (storyboarding, affinity mapping, lotus blossom) (v1.32) |
| **131. Pathway Router** | 5 educational pathways with signal-word matching, mid-session adaptive resequencing on energy/saturation/user signals (v1.32) |
| **132. 8 Specialized Agents** | Facilitator (assessment, transitions, energy), Ideator, Questioner, Analyst (SCAMPER/Hats), Mapper, Persona (constructive figures), Critic (Converge-only), Scribe (always-on capture) (v1.32) |
| **133. SessionBus & Integration** | 4-loop filesystem message router, 18 safety-critical tests, 3 E2E pathway tests, chipset YAML with 4 activation profiles (v1.32) |
| **134. OpenStack Skills** | 19 GSD skills: 8 core services (keystone/nova/neutron/cinder/glance/swift/heat/horizon), 6 ops (monitoring/backup/security/networking-debug/capacity/kolla-ansible-ops), 4 doc/methodology, kolla-ansible deployment (v1.33) |
| **135. Mission Crews** | 3 crew configurations (31 agents): Deployment (12 roles), Operations (8 roles), Documentation (8 roles) with Scout/Patrol/Squadron profiles and crew handoff (v1.33) |
| **136. Cloud ASIC Chipset** | Complete chipset.yaml: 19 skills, 31 agents, 9 communication loops, pre/post-deploy evaluation gates, budget enforcement (118/118 validation checks) (v1.33) |
| **137. NASA SE Documentation** | 7-chapter sysadmin guide mapping SE phases, 8-service operations manual (80 procedures), 44 runbooks with dual indexes, 3-tier reference library, cross-cloud translation (v1.33) |
| **138. V&V Infrastructure** | Requirements verification matrix (55 reqs), NPR 7123.1 compliance matrix, 22 safety-critical tests, E2E deployment and user scenario verification scripts (v1.33) |
| **139. Cloud-Ops Modules** | TypeScript dashboard panel, config staging pipeline, deployment observer, git commit rationale formatter, 3-tier knowledge loader (216 tests) (v1.33) |
| **140. Documentation Canonical Source** | 158 markdown files as canonical source of truth with YAML frontmatter schema, narrative spine connecting 5 educational layers, 7 gateway documents, 4 extractable templates (educational-pack, career-pathway, AI-learning-prompt, mission-retrospective), site architecture with Astro SSG and 5-phase WordPress migration (v1.34) |
| **141. Mathematical Foundations Engine** | 451 typed primitives across 10 domains with 106 dependency edges, 8 mathematical engines (graph, path-finder, classifier, navigator, composition, proof, verification, property-checkers), 58 composition rules (v1.35) |
| **142. sc:learn Knowledge Ingestion** | Generalized knowledge pipeline: acquire (PDF/md/docx/txt/epub/zip/tgz/GitHub) → sanitize (STRANGER-tier, 6 attack categories) → HITL gate → analyze → extract → deduplicate → generate skills/agents/teams → report with provenance (v1.35) |
| **143. sc:unlearn Reversible Sessions** | Reversible knowledge ingestion with changeset manager, graph integrity validation, and skill regeneration (v1.35) |
| **144. Citation Extraction & Resolution** | Multi-format citation extractor (8 regex patterns: APA, narrative, numbered, DOI, ISBN, URL, bibliography, prose), 6-adapter resolution cascade (CrossRef, OpenAlex, NASA NTRS, GitHub, Internet Archive, generic web) with confidence thresholds (v1.36) |
| **145. Bibliography Generation** | 5 output formats (BibTeX, APA 7th, Chicago 17th, MLA 9th, custom Mustache), attribution reports with 5 claim categories, provenance chain tracking with bidirectional queries, integrity auditor (v1.36) |
| **146. Complex Plane Skill Positioning** | SkillPosition (θ angle, r radius) replacing linear model, tangent-line activation with geometric/semantic weight blending (60/40), Score stage integration preserving 5 downstream pipeline stages (v1.37) |
| **147. Angular Promotion Pipeline** | 7-check AngularPromotionEvaluator (direction, adjacency, step, evidence, stability, velocity, F1/MCC), AngularRefinementWrapper with content direction analysis, CONSTRAINT_MAP preserving 3-correction/7-day/20% rules (v1.37) |
| **148. Chord Detection & Euler Composition** | ChordDetector with arc/savings/frequency filtering, EulerCompositionEngine (complex multiplication r1·r2·e^(i(θ1+θ2))), quality assessment with action recommendation, co-activation geometric gate (v1.37) |
| **149. Plane Metrics & Migration** | Versine distribution, exsecant reach, velocity warnings, `skill-creator plane-status` CLI (--json/--snapshot/--detail), SkillMigrationAnalyzer with dry-run/force/idempotent modes (v1.37) |
| **150. OS-Level Agent Sandbox** | Bubblewrap (Linux) and Seatbelt (macOS) per-agent profiles for 4 agent types, SSH key/credential directory denial, escape prevention (nsenter/unshare/chroot/proc blocking) (v1.38) |
| **151. Zero-Knowledge Credential Proxy** | SecretString with zeroize-on-drop, Unix socket server (mode 0600), domain allowlist (deny-by-default), API key injection without credentials entering sandbox, GNOME Keyring/macOS Keychain native stores (v1.38) |
| **152. CVE Staging Security Scanner** | 8 pattern detectors (SEC-001 through SEC-008): settings.json hooks, API redirect, hook injection, MCP risk, sandbox escape, SSH key references, credential exfiltration, base64 obfuscation, agent-proof quarantine (v1.38) |
| **153. Security Dashboard** | Shield indicator at all 5 magic levels, 4 shield states (secure/attention/breach-blocked/inactive), critical events bypass magic filter, event timeline, security operations view (v1.38) |
| **154. IPC Foundation** | 29 Zod+serde event type schemas with TS-Rust JSON parity, 9 Tauri command stubs, desktop IPC wrappers, .planning/ directory scaffold (v1.39) |
| **155. Rust SSE API Client** | Anthropic Messages API streaming via SSE, secure key management (env/keychain/encrypted file), exponential backoff retry with 429 compliance, conversation history persistence (v1.39) |
| **156. CLI Chat & READY Prompt** | Terminal-styled webview chat with READY. sequence, streaming render from IPC deltas, command history, auto-scroll, LED status panel, XSS prevention (v1.39) |
| **157. Magic Verbosity Control** | 5-level system (Full Magic L1 to No Magic L5), event visibility map for 29 event types, chat passthrough, JSON persistence, Settings recalibrate panel with live preview (v1.39) |
| **158. Service Launcher** | 7-service dependency graph with Kahn's topological sort, ordered startup with dependency validation, 5s health checks (3→Degraded, 5→Failed), LED state machine, graceful shutdown (v1.39) |
| **159. Staging Intake** | Filesystem watcher with atomic move-to-processing, hygiene pipeline (YAML code execution, path traversal, zip extraction), quarantine routing, debrief collector with calibration ratios (v1.39) |
| **160. Self-Improvement Lifecycle** | RETROSPECTIVE.md template generator, changelog watch with LEVERAGE/PLAN/WATCH classification, calibration delta computation, observation harvester, prioritized action generator (v1.39) |
| **161. PDF Extraction Pipeline** | pdftotext-based extraction with chapter/part detection, LaTeX math preservation, TikZ diagram cataloging, MusiXTeX tagging, exercise/build-lab extraction, 8000-token chunk segmentation, JSONL manifest (v1.40) |
| **162. Ingestion Harness** | Atomic checkpoint write/read with SHA-256 integrity hash, multi-session resume with validation, per-chapter metrics (tokens, concepts, math density), dashboard bridge writing JSON to console outbox (v1.40) |
| **163. Dual-Track Learning** | Track A (Parts I-V) and Track B (Parts VI-X) concept extraction with regex-based detection, complex plane positioning via PART_ANGULAR_REGIONS, progressive depth tracking, ecosystem cross-referencing, token budget enforcement (v1.40) |
| **164. Three-Track Verification Engine** | Track A concept coverage audit via Jaccard similarity, Track B cross-document consistency via antonym-pair detection, Track C eight-layer-to-ten-part progression mapping, 8-type gap taxonomy, 4-level severity (v1.40) |
| **165. Refinement & Reporting** | Knowledge patch generator (proposals only, requiresReview=true), improvement ticket generator with reproduction steps, skill refiner with complex plane positions, 11-section dogfood report builder, safety validator (v1.40) |
| **166. Safety Validation** | Bounded learning enforcement (≤20% via Levenshtein), checkpoint integrity (SHA-256), no auto-application (requiresReview=true on all patches), regression gate (16,795 tests, 0 failures) (v1.40) |
| **167. Slim CLAUDE.md Architecture** | 417→46 line CLAUDE.md migration with 4 auto-activating skills (gsd-workflow, skill-integration, session-awareness, security-hygiene), 3 scoped subagent definitions (executor, verifier, planner), reference document architecture (v1.41) |
| **168. Git Workflow Intelligence** | Git-aware pattern detection from commit history, progressive disclosure git workflow skill, CLI command registration, coverage reporting with @vitest/coverage-v8 (v1.42) |
| **169. Gource Visualization Pipeline** | Configuration profiles for repository history animation, batch processing scripts, avatar/caption management, BATS test suites, deployment documentation (v1.43) |
| **170. PyDMD Knowledge Dogfood** | sc:learn pipeline dogfood against PyDMD scientific computing library, concept mapping to knowledge graph, tutorial replay engine, accuracy checker, observation bridge, dogfood report template (v1.44) |
| **171. Agent-Ready Static Site Generator** | Custom SSG for tibsfox.com with markdown→HTML pipeline, Mustache template engine, agent discovery layers (llms.txt, llms-full.txt, AGENTS.md, Schema.org JSON-LD), client-side search, citation integration, WordPress MCP sync (v1.45) |
| **172. Upstream Intelligence Pack** | Upstream change monitoring with 5 agents (SENTINEL, ANALYST, TRACER, PATCHER, HERALD), 3 team topologies, channel monitors for Anthropic API/SDK, append-only persistence with rollback, 50-event historical test corpus (v1.46) |
| **173. Fluid Systems Engineering** | Darcy-Weisbach and Hazen-Williams pressure drop calculations, ASHRAE TC 9.9 thermal guidelines, CDU selection, pump curve modeling with system curve intersection, glycol concentration effects (v1.48) |
| **174. Power Systems Engineering** | NEC 220/310/690 calculations, transformer K-factor derating, UPS runtime with battery aging, PDU whip scheduling, BESS capacity planning, DC distribution voltage drop, solar PV string sizing (v1.48) |
| **175. Thermal Engineering** | Conduction/convection/radiation with composite R-values, LMTD and effectiveness-NTU heat exchanger methods, PUE/TUE/WUE efficiency metrics, cooling load estimation, free cooling economizer hours (v1.48) |
| **176. Blueprint Engine** | P&ID generation with 50 ISA-5.1 symbols, SLD with 30 IEEE C2-2023 symbols, floor plan layout with hot/cold aisle containment, isometric piping views, SVG output with DXF export, parametric symbol helpers (v1.48) |
| **177. Dimensional Analysis** | Physical unit tracking (SI/Imperial), tolerance stack-up (RSS/worst-case), Buckingham Pi dimensional homogeneity, interference checking between piping/conduit/structural, NEC 392 cable tray fill (v1.48) |
| **178. Simulation Bridge** | OpenFOAM case file generation, ngspice netlist generation, React artifact output, Minecraft redstone mapping, Factorio blueprint patterns, FreeCAD macro generation — input generation only, correct scope boundary (v1.48) |
| **179. Construction Documents** | BOM generation with quantity takeoff, installation sequence planning with dependencies, commissioning checklists (ASHRAE Guideline 0, NETA ATS), O&M manual generation, spare parts with A/B/C criticality (v1.48) |
| **180. Creative Pipeline** | Blender bpy script generation for 3D walkthroughs, ffmpeg assembly for construction timelapse, social media export templates (LinkedIn, Instagram carousel), render preset management (v1.48) |
| **181. Safety Warden Agent** | 3 operating modes (annotate/gate/redirect), mandatory PE disclaimer on all engineering outputs, non-bypassable structural architecture, HITL gate for calculation application, IEC/NEC/ASHRAE citation on every output (v1.48) |
| **182. Physical Infrastructure Chipset** | Router topology with Architect (Opus) entry point, 4 Sonnet specialist agents (Fluid, Power, Thermal, Blueprint), Safety Warden crosscut, 3 team topologies (domain-specialist, cross-domain, full-build), 3 E2E integration pipelines (v1.48) |
| **183. Usage Event Emission & JSONL EventStore** | UsageEvent types with lifecycle tracking (load, activate, deactivate, error), append-only JSONL EventStore with session-aware event correlation, automatic emission integrated into skill lifecycle (v1.49.13) |
| **184. Skill Health Pattern Detector** | UsagePatternDetector identifying 7 skill health patterns: load-never-activate, correction-magnet, single-session-spike, co-activation cluster, score-drift, seasonal, session-spread (v1.49.13) |
| **185. Adaptive Feedback Pipeline** | ScoreAdjuster with bounded +/-20% multipliers per adjustment cycle, CachePromoter for automated cache tier promotion, sc:digest integration (Step 4.5), 90-day retention enforcement with automatic event pruning (v1.49.13) |
| **186. Multi-Ecosystem Dependency Auditor** | Manifest parsing for 5 ecosystems (npm, PyPI, conda, crates.io, RubyGems), OSV.dev vulnerability scanning with severity mapping, shared rate-limiter (<=30 req/60s), incremental scanning of changed dependencies only (v1.49.14) |
| **187. Dependency Health Diagnostician** | 6-class health classification (healthy, aging, stale, abandoned, deprecated, vulnerable), ecosystem-aware thresholds, Python compatibility matrix, conflict detection, P0-P3 severity ranking (v1.49.14) |
| **188. Alternative Discoverer** | SuccessorDetector, ForkFinder, EquivalentSearcher, InternalizationFlagger with structured evidence and confidence scores, relationship type classification (v1.49.14) |
| **189. HITL Dependency Resolver** | Backup before every manifest change, dry-run verification, human-in-the-loop approval gate, one dependency per resolution (blast radius control), instant rollback from backup (v1.49.14) |
| **190. Code Absorber with Oracle Testing** | Criteria gate (<=500 LOC, stable API, pure functions), oracle testing against 10,000+ generated cases, gradual call-site replacement (<=20% tranches), hard block on crypto/parsers/protocols/compression (v1.49.14) |
| **191. Cross-Project Health Learning** | Append-only health.jsonl with mandatory provenance, StagingHealthGate for pipeline integration, PatternLearner with pre-emptive warnings after 5+ projects flag same dependency (v1.49.14) |
| **192. Model Abstraction Layer** | Chip drivers for OpenAI-compatible and Anthropic endpoints, ChipRegistry with CLI integration, Zod-validated chipset.json configuration schemas (v1.49.15) |
| **193. Multi-Model Evaluation** | Benchmark schema with grader calibration, model-aware grading with tier detection, capability profiling, threshold configuration (v1.49.15) |
| **194. MCP Mesh Infrastructure** | LLM Wrapper with per-chip queuing and timeout handling, Mesh Discovery with node health monitoring and stale eviction, DACP Transport with provenance tracking and fidelity-adaptive compression (v1.49.15) |
| **195. Mesh Orchestration & Failover** | Pure scoring functions (scoreNode, rankNodes, selectWithFallback), MeshCoordinator with dispatch, fan-out, pipeline, and automatic failover, VTM wave planner and multi-model optimizer (v1.49.15) |
| **196. Cross-Mesh Context Preservation & Skill Packaging** | Context preservation across mesh boundaries with transcript summarization, mesh git worktrees with proxy committer, skill manifest and packager with tested_models and mesh_hints, Skill Creator MCP Server with 8 pipeline tools (v1.49.15) |
| **197. Skill Lifecycle & MCP Tool Handlers** | SkillLifecycleResolver with draft/tested/graded state derivation, SkillCreatorMCP server with 10 real tool handlers (compare, analyze, optimize, package, benchmark), pipeline wiring (v1.49.16) |
| **198. Task Execution Engine** | TaskTracker with lifecycle state machine and event emission, MeshExecutor with wave, pipeline, and parallel dispatch modes (v1.49.16) |
| **199. Model Capability Classification & Calibration** | CapabilityClassifier and LimitationRegistry, extended ModelAwareGrader with grading context and calibrated hints, CalibrationStore for eval persistence (v1.49.16) |
| **200. Optimization Driver with Convergence Detection** | ConvergenceDetector and VariantGenerator, OptimizationDriver with convergence loop and variant forking (v1.49.16) |
| **201. Muse Schema Foundation** | Muse schema with 6 system definitions and Zod validator, wired into chipset barrel exports, CI guard to block .planning/ files from being pushed (v1.49.16) |
| **202. Cartridge Packaging Format** | New composition mechanism for knowledge packs — cartridge types, loader, validator, registry, packer/unpacker, integration with existing pack system (v1.49.17) |
| **203. Audio Engineering Educational Pack** | 36 concepts across 6 domains (synthesis, physics of sound, consoles & mixing, DJ culture, MIDI & protocols, production), 32 citations from foundational literature, college enrichments for music and physics departments (v1.49.17) |
| **204. Hardware Infrastructure Pack** | 5 hardware tiers (edge, desktop, workstation, server, cloud) with node profiles and type system, college enrichments for electronics and engineering departments, mesh architecture mapping (v1.49.17) |
| **205. Space Between Cartridge & Muse Vocabulary** | 31 integrated concepts bridging audio, hardware, and mathematics, 122-entry foundation vocabulary index for the creative insight engine (Cedar), cross-domain concept bridging, integration tests (v1.49.17) |
| **206. Observatory Shell & Telescope** | Navigation, state management, dark theme, unified telescope view with 4 parallel chains, learner progression across all 8 mathematical foundations simultaneously (v1.49.18) |
| **207. 8 Mathematical Foundation Wings** | One wing per foundation (Unit Circle, Pythagorean, Trigonometry, Vector Calculus, Set Theory, Category Theory, Information Theory, L-Systems), each with 6-phase content (wonder, see, touch, understand, connect, create) (v1.49.18) |
| **208. Interactive Canvas & Nature Simulations** | Canvas system with RAF loop, FPS tracking, pointer normalization, 5 nature simulations (set visualizer, functor bridge, tree growth, magnetic field), generative art tools (v1.49.18) |
| **209. Garden Creative Workshop** | Art canvas, music studio (Tone.js), L-system editor, reflective journal — create with mathematics before learning notation (v1.49.18) |
| **210. Narrative Layer & Foundation Infrastructure** | 8 wonder stories (zero mathematical notation), 8 literary bridges, 20 reflection prompts, foundation registry (1,489 lines, 6 phases each), connection graph with BFS pathfinding, progression engine, Wonder Warden with soft/hard phase gating (v1.49.18) |
| **211. Gastown Chipset Definition & Validator** | Complete YAML with agent topology, communication channels, dispatch config, 4-stage validator (schema, token budget, topology, channels), JSON Schema draft-07 (v1.49.19) |
| **212. 12 Multi-Agent Orchestration Skills** | Mayor, Polecat, Witness, Refinery, Mail, Nudge, Hook, Sling, Done, Runtime HAL, GUPP, Beads — absorbed from steveyegge/gastown into GSD ecosystem (v1.49.19) |
| **213. StateManager & Integration Documentation** | StateManager with atomic writes and crash recovery, 18 TypeScript interfaces, barrel export wiring, 10-document integration guide in docs/gastown-integration/ covering architecture, security model, trust boundaries, agent topology, and GSD workflow (v1.49.19) |
| **214. Documentation Consolidation** | Comprehensive documentation audit (22 findings across 158+ files), 6 resolved during consolidation (wrong GitHub URL, stale core concepts, missing wave execution docs, missing FEATURES entries, missing RELEASE-HISTORY entries, missing Gastown README links), payload-agnostic execution pattern documented (v1.49.20) |
| **215. Documentation Reflections** | 16 remaining audit findings resolved: stale metrics in about.md and index.md updated to 85 milestones, placeholder files across 5 doc directories replaced with real content, stale package name fixed, stale version output fixed, filesystem-contracts supersession notice, GSD Orchestrator cross-reference to Mayor-coordinator, WordPress migration status corrected, OpenStack educational context headers (v1.49.20.1) |
| **216. Image to Mission Pipeline** | Creative translation pipeline: observation engine (4-layer visual analysis), context integrator (freeform parser with process insight), connection engine (cross-image synthesis), parameter extractor (color/geometry/material/feel), dual translation engines (code: Canvas/React/Three.js/CSS; design: SVG/palette/layout), build generator (atomic steps with philosophy), transmission packager (5 self-containment checks), pipeline bridge (complexity scoring 0-12, routing to vision-to-mission), 14 Zod schemas, image-to-mission skill definition, 253 tests (v1.49.21) |
| **217. PNW Research Browser Series** | Static research browsers for Pacific Northwest biodiversity: Columbia Valley Rainforest (`www/tibsfox/com/Research/COL/`, 8 research documents), Cascade Range (`www/tibsfox/com/Research/CAS/`, 10 research documents across elevation zones), PNW Living Systems (`www/tibsfox/com/Research/ECO/`, dual mission packs with LaTeX sources), PNW Gardening (`www/tibsfox/com/Research/GDN/`), series master index with deep cross-linking, client-side markdown rendering, no external dependencies (v1.49.22) |
| **218. GPU-Accelerated Math Co-Processor** | 5 computational chips (Algebrus, Fourier, Statos, Symbex, Vectora) with CUDA GPU acceleration and NumPy CPU fallback, 18 MCP tools via Model Context Protocol, config-driven chip activation, VRAM monitoring, CUDA stream isolation, chipset YAML definition, 125 tests (v1.49.22) |
| **219. SYMBEX JIT Compiler** | AST-based Python-to-CUDA C translator with NVRTC runtime compilation, kernel cache, operator precedence and type inference, 48 JIT-specific tests (v1.49.23) |
| **220. Full GPU Acceleration Paths** | cuSOLVER (dense linear solve, SVD, eigenvalue decomposition), cuFFT (FFT/IFFT), cuRAND (Monte Carlo with configurable distributions), completing all 5 chip GPU backends (v1.49.23) |
| **221. Unison Language Research & Chipset** | Deep research study (5 modules + synthesis + verification, 28,693 words, 28 sources) at `www/tibsfox/com/UNI/`, translation chipset with 6 skills, 4 agents, and translation maps for Haskell, TypeScript, Rust, and Erlang equivalences (v1.49.23) |
| **222. Dashboard MCP Integration** | MathBridge polling MCP server for live math co-processor data, operation tracing, 16 tests (v1.49.23) |
| **223. Building Construction Mastery** | PNW building science research browser (`www/tibsfox/com/Research/BCM/`) with 6 trade disciplines at 5 audience levels: structural systems (ASCE 7-22 seismic), electrical systems (NEC 2023), plumbing & mechanical (UPC 2021), building envelope, codes & standards (dual OR/WA mapping), educational frameworks (ABET alignment), 119 safety callouts, 12 research files, 680 KB, 48/48 verification tests (v1.49.24) |
| **224. Smart Home & DIY Electronics** | Complete electronics curriculum browser (`www/tibsfox/com/Research/SHE/`) with 36 hands-on projects across 6 tiers, 22 sensors, 14 actuators, 10 communication protocols (MQTT, Zigbee, Z-Wave, Thread, Matter, Wi-Fi, BLE, LoRa, I2C, SPI), Home Assistant/ESPHome/ESP32 platform coverage, NEC safety integration, 56 safety callouts, 12 research files, 532 KB, 33/33 verification tests (v1.49.24) |

## Version History

| Version | Key Features |
|---------|--------------|
| **v1.0** | Core skill management, pattern observation, learning loop, agent composition |
| **v1.1** | Semantic conflict detection, activation scoring, local embeddings via HuggingFace |
| **v1.2** | Test infrastructure, activation simulation, threshold calibration, benchmarking |
| **v1.3** | Documentation overhaul, official format specification, getting started guide |
| **v1.4** | Agent Teams: team schemas, storage, validation, CLI commands, GSD workflow templates |
| **v1.5** | Pattern Discovery: session log scanning, tool sequence extraction, DBSCAN clustering, draft generation |
| **v1.6** | 34 cross-domain examples (20 skills, 8 agents, 3 teams), local installation, beautiful-commits skill |
| **v1.7** | GSD Master Orchestration Agent: dynamic discovery, intent classification, lifecycle coordination, verbosity/HITL gates, persistent work state, session continuity, skill workflows, roles, bundles, inter-skill events |
| **v1.8** | Capability-Aware Planning + Token Efficiency: skill pipeline architecture, per-agent token budgets, capability manifests, phase capability declarations, skill injection, cache-aware ordering, research compression, model-aware activation, collector agents, parallelization advisor |
| **v1.8.1** | Audit Remediation: test infrastructure fixes, type safety improvements, CLI validation, error handling, dependency validation, security hardening, code refactoring, cache invalidation |
| **v1.9** | Ecosystem Alignment & Advanced Orchestration: spec-aligned skill generation, progressive disclosure, 5-platform portability, evaluator-optimizer with A/B testing, MCP-based distribution, router/map-reduce topologies, session save/restore/handoff, agentic RAG with corrective refinement, quality-of-life CLI improvements |
| **v1.10** | Security Hardening: path traversal prevention, YAML safe deserialization, JSONL integrity (checksums, rate limiting, anomaly detection, compaction), discovery safety (secret redaction, allowlist/blocklist, deny list), learning safety (drift tracking, contradiction detection), team message sanitization, config validation, inheritance chain validation, file integrity monitoring, hook error boundaries, SECURITY.md, CI pipeline |
| **v1.11** | GSD Integration Layer: integration config with per-feature toggles, idempotent install script with --uninstall, POSIX shell post-commit hook, 6 slash commands (/sc:start, status, suggest, observe, digest, wrap), 4 wrapper commands (/wrap:execute, verify, plan, phase with smart routing), passive monitoring (plan-vs-summary diffing, STATE.md transitions, ROADMAP.md structural diff) |
| **v1.12** | GSD Planning Docs Dashboard: markdown-to-HTML generator for `.planning/` artifacts, 5 dashboard pages with dark theme, JSON-LD structured data, Open Graph meta tags, incremental builds with SHA-256 hashing, auto-refresh with scroll preservation, GSD slash command |
| **v1.12.1** | Live Metrics Dashboard: three-tier sample rate engine (hot/warm/cold), live session pulse, phase velocity analytics, planning quality scores, historical trends, CSS-only visualizations, 460 dashboard tests |
| **v1.13** | Session Lifecycle & Workflow Coprocessor: message stack (push/pop/poke/drain), session management via tmux, recording/playback with 14 metrics, Pipeline List coprocessor (WAIT/MOVE/SKIP), Offload engine for script promotion, Team-as-Chip framework (Agnus/Denise/Paula/Gary), Exec kernel with prioritized scheduling, Pipeline learning from observations, 1057 tests |
| **v1.14** | Promotion Pipeline: execution capture with SHA-256 hashes, three-tier determinism analysis, weighted promotion scoring (determinism 40%/frequency 35%/savings 25%), script generation with dry-run, F1/MCC gatekeeper, drift monitoring with auto-demotion, bidirectional lineage tracking, 3 dashboard collectors |
| **v1.15** | Live Dashboard Terminal: Wetty terminal config with Zod validation, process management with health checks, tmux session binding with auto-detection, themed iframe panel with offline fallback, unified DevEnvironmentManager launcher |
| **v1.16** | Dashboard Console & Milestone Ingestion: filesystem message bus with Zod-validated JSON envelopes, HTTP helper with path traversal prevention, drag-and-drop upload zone, inbox checking at lifecycle boundaries, 5 interactive question types, console dashboard page with hot-configurable settings |
| **v1.17** | Staging Layer: 5-state filesystem pipeline (inbox/checking/attention/ready/aside), 11-pattern hygiene engine, trust-aware reporting with familiarity tiers, smart intake with clarity routing, resource analysis with skill cross-reference, derived knowledge checking with provenance chains, 7-state staging queue with audit log, queue pipelining with pre-wiring |
| **v1.18** | Information Design System: CSS design system (6 domain colors, 4 signal colors, typography tokens), persistent gantry status strip, 6 SVG entity shapes with dual encoding, subway-map topology with bezier edges and click-to-detail, activity feed with Unicode indicators, budget gauge with threshold transitions, domain-prefixed identifiers (F-1, B-1.api, T-1:rcp) |
| **v1.19** | Budget Display Overhaul: LoadingProjection separating installed from loadable with tier-based selection, two-section CLI status with proportional sizing and color-coded budget bar, JSON output mode, dashboard gauge with deferred tooltip and over-budget clamping, per-profile cumulative budgets in config, dual-dimension budget history with migration |
| **v1.20** | Dashboard Assembly: unified CSS pipeline (18 component style modules), 4 data collectors (topology, activity, budget-silicon, staging queue), console page assembly as 6th dashboard page |
| **v1.21** | GSD-OS Desktop Foundation: Tauri v2 desktop shell with Rust backend, WebGL CRT shader engine, 32-color indexed palette with 5 presets, native PTY terminal with xterm.js, Amiga Workbench window manager, boot sequence and calibration wizard, 636 tests |
| **v1.22** | Minecraft Knowledge World: local cloud infrastructure (PXE/kickstart/VM provisioning), Minecraft Fabric server with Syncmatica, platform portability (3 package managers, 3 hypervisors), Amiga emulation (FS-UAE/AROS), chipset formalization (20 skills, 10 agents, 5 teams), operational maturity (backups, monitoring, runbooks) |
| **v1.23** | Project AMIGA: AMIGA mission infrastructure (MC-1/ME-1/CE-1/GL-1 with 4 ICDs), Apollo AGC Block II simulator (38 instructions, bank-switched memory, interrupts, I/O), AGC Executive/Waitlist/BAILOUT, DSKY interface with Executive Monitor, AGC development tools (assembler, debugger, disassembler, rope loader), AGC curriculum (11 chapters, 8 exercises), RFC Reference Skill (3-agent system, 57-RFC index) |
| **v1.24** | GSD Conformance Audit & Hardening: 336-checkpoint conformance matrix against 18 vision documents, full 4-tier audit (T0/T1/T2/T3) with 211 pass + 125 amendments, zero-fail conformance achieved |
| **v1.25** | Ecosystem Integration: dependency DAG (20 nodes, 48 edges), EventDispatcher specification, 4-tier dependency philosophy, integration test strategy, partial-build compatibility matrix with 99 known-issues |
| **v1.26** | Aminet Archive Extension Pack: INDEX infrastructure (~84K packages), mirror state + download engine, full-text search + collections, pure TypeScript virus scanner (52 signatures), LhA/LZX extraction, FS-UAE integration, GSD-OS desktop panel with 5-agent pipeline |
| **v1.27** | GSD Foundational Knowledge Packs: 35 knowledge packs (15 Core Academic, 10 Applied, 10 Specialized) with complete vision/modules/activities/assessment/resources, GSD-OS dashboard (browser, search, detail view, skill tree, progress tracking), skill-creator integration (observation hooks, AMIGA bridge, 4-pattern detection, pathway adaptation, activity scaffolding, approach promotion) |
| **v1.28** | GSD Den Operations: filesystem message bus (8 priorities, ISA encoding), 10-position staff topology (5 divisions), 4 topology profiles (Scout/Patrol/Squadron/Fleet), HALT/CLEAR protocol, Sentinel recovery matrix, Chronicler audit trail, chipset YAML reproducibility, 675 tests |
| **v1.29** | Electronics Educational Pack: MNA circuit simulator (7 components, DC/AC/transient), digital logic simulator (8 gates, flip-flops, timing), safety warden (IEC 60449, positive framing), learn mode (3-level H&H), 15 modules across 4 tiers with 77 labs, 5 specialized engines (DSP/GPIO/PLC/solar/PCB), 10,707 tests |
| **v1.30** | Vision-to-Mission Pipeline: Zod-validated VTM types, vision parser with archetype classification, research compiler with tiered chunking, mission assembler, wave planner with graph coloring, model assignment (Opus/Sonnet/Haiku), cache optimizer, test plan generator, 7-template system, pipeline orchestrator with enrichment, 679 tests |
| **v1.31** | GSD-OS MCP Integration: Rust MCP Host Manager (server lifecycle, tool routing, trace emission), Gateway Server (19 tools, Streamable HTTP, scope enforcement), MCP templates (server/host/client scaffolds), Agent Bridge (SCOUT/VERIFY servers, EXEC client), security pipeline (hash verification, trust lifecycle, staging gates), Blueprint Editor blocks with wiring engine, Trace Panel, Security Dashboard, 838 tests |
| **v1.32** | Brainstorm Session Support: 8 specialized agents (Facilitator, Ideator, Questioner, Analyst, Mapper, Persona, Critic, Scribe), 16 techniques across 4 categories, 5 educational pathways with adaptive routing, Osborn's rules engine with architectural Critic gate, SessionBus 4-loop message router, session lifecycle (Explore→Diverge→Organize→Converge→Act), artifact generation, 321 tests |
| **v1.33** | GSD OpenStack Cloud Platform (NASA SE Edition): 19 skills (8 core + 6 ops + 4 doc + kolla-ansible), 3 crew configurations (31 agents) with Scout/Patrol/Squadron profiles, ASIC chipset (19 skills, 31 agents, 9 comm loops, evaluation gates), 7-chapter NASA SE sysadmin guide, 8-service ops manual (80 procedures), 44 runbooks with dual indexes, 3-tier reference library, V&V with NPR 7123.1 compliance, 22 safety-critical tests, cloud-ops TypeScript modules (dashboard, staging, observation, git, knowledge), 216 tests |
| **v1.34** | Documentation Ecosystem Refinement: 158 markdown files as canonical source of truth, narrative spine with 5 educational layers, 7 gateway documents, 4 extractable templates, Astro SSG site architecture with 5-phase WordPress migration |
| **v1.35** | Mathematical Foundations Engine + sc:learn: 451 primitives across 10 domains, 8 mathematical engines, 58 composition rules, generalized knowledge ingestion pipeline (acquire/sanitize/HITL/analyze/extract/deduplicate/generate/report), sc:unlearn reversible sessions, 631 tests |
| **v1.36** | Citation Management & Source Attribution: multi-format citation extractor (8 patterns), 6-adapter resolution cascade, 5 bibliography formats, attribution reports, provenance chain tracking, citation chipset (6 skills, 4 agents), 280 tests |
| **v1.37** | Complex Plane Learning Framework: SkillPosition (angle, radius) replacing linear model, tangent-line activation, angular promotion pipeline, chord detection, Euler composition, versine/exsecant metrics, plane migration system, 446 tests |
| **v1.38** | SSH Agent Security: OS-level bubblewrap/Seatbelt sandboxing, zero-knowledge credential proxy, CVE staging security scanner (8 detectors), per-agent isolation via git worktrees, security dashboard, Bootstrap Phase 0, 260 tests |
| **v1.39** | GSD-OS Bootstrap & READY Prompt: 29 IPC event types, Rust SSE API client, CLI chat with READY. sequence, 5-level magic verbosity, 7-service launcher, staging intake, self-improvement lifecycle, 517 tests |
| **v1.40** | sc:learn Dogfood Mission: PDF extraction pipeline, ingestion harness with checkpoint resume, dual-track learning (Parts I-V / VI-X), three-track verification engine, refinement and reporting, bounded learning safety validation, 362 tests |
| **v1.41** | Claude Code Integration Reliability: slim CLAUDE.md (417→46 lines), 4 auto-activating skills, 3 subagent definitions, install script migration, reference document architecture |
| **v1.42** | SC Git Support: git workflow intelligence, pattern detection from commits, progressive disclosure skill, CLI integration, coverage reporting, 272 tests |
| **v1.43** | Gource Visualization Pack: repository animation pipeline, configuration profiles, batch processing, avatar/caption systems, 46 BATS tests |
| **v1.44** | SC Learn PyDMD Dogfood: PyDMD knowledge extraction, concept mapping, tutorial replay engine, accuracy checker, observation bridge, dogfood report, 284 tests |
| **v1.45** | Agent-Ready Static Site: custom SSG with agent discovery (llms.txt, AGENTS.md, JSON-LD), template engine, search, citations, WordPress sync, deployment, 243 tests |
| **v1.46** | Upstream Intelligence Pack: 5 monitoring agents, 3 team topologies, channel monitors, persistence with rollback, 50-event test corpus, pipeline orchestrator, 206 tests |
| **v1.47** | Holomorphic Dynamics Educational Pack: 10-module pack (HD-01 through HD-10) covering iteration, fixed points, Mandelbrot/Julia sets, cycles, topology, deep learning dynamics, skill-creator dynamics, DMD, and Koopman theory. Complex arithmetic library, fractal renderer, standard DMD + 4 variants (DMDc, mrDMD, piDMD, BOP-DMD), EDMD with dictionary lifting, SkillDynamicsExtended bridge, skill dynamics classification with Fatou/Julia domains, 209 tests |
| **v1.48** | Physical Infrastructure Engineering Pack: fluid systems (Darcy-Weisbach, Hazen-Williams, pump curves), power systems (NEC 220/310/690, transformers, UPS, solar PV, BESS), thermal engineering (LMTD, e-NTU, PUE/TUE/WUE), blueprint engine (50 ISA-5.1 + 30 IEEE symbols, P&ID/SLD/floor plan/isometric, SVG/DXF), dimensional analysis (units, tolerances, Buckingham Pi), simulation bridge (OpenFOAM, ngspice, FreeCAD), construction documents (BOM, commissioning, O&M), creative pipeline (Blender, ffmpeg), Safety Warden (annotate/gate/redirect, PE disclaimer, non-bypassable), router topology chipset (Architect + 4 specialists, 3 teams), 3 E2E integration pipelines, educational bridge (Minecraft/Factorio), 401 tests |
| **v1.49** | Deterministic Agent Communication Protocol: 26 Zod schemas + 9 JSON Schemas for DACP types, directory-based bundle format (manifest.json + intent.md + data/ + code/ with .complete markers), adaptive fidelity model (Level 0-3, 95% accuracy), DACP Assembler (three-part bundle composition from skill library), DACP Interpreter (8-stage validation, Object.freeze no-auto-execute, provenance enforcement), retrospective analyzer (drift scoring, pattern detection, cooldown enforcement), script catalog + schema library (function-type indexing, provenance, EMA tracking), 5 bundle templates, filesystem bus integration with .msg fallback, handoff dashboard panel, 5 CLI commands, 8/8 safety-critical tests, 263+ verification tests |
| **v1.49.24** | Building Construction & Smart Home Electronics: BCM research browser (6 trades, NEC 2023, UPC 2021, ASCE 7-22, dual OR/WA code mapping, 119 safety callouts, 48/48 tests), SHE research browser (36 projects, 6 tiers, 10 protocols, 22 sensors, 14 actuators, 56 safety callouts, 33/33 tests), PNW series expanded to 6 projects (66 files, 3.8 MB, 124 sources) |
| **v1.49.23** | Unison Language & Math Co-Processor Completion: cuSOLVER/cuFFT/cuRAND GPU paths, SYMBEX JIT compiler (AST-based Python→CUDA C, NVRTC, kernel cache), Unison Language deep research (5 modules, 28,693 words, 28 sources), Unison translation chipset (6 skills, 4 agents, 4 language maps), dashboard MCP wiring, PNW research collection, Aspen muse |
| **v1.49.25** | PNW Wildlife Survey: Wings of the Pacific Northwest (AVI) documenting 250+ bird species across 7 ecoregions (19 files, 2.1 MB, 88 sources, 38/38 verification, 8/8 safety-critical), Fur Fin & Fang (MAM) documenting 169+ mammal species across 8 ecoregion zones including marine (14 files, 1.1 MB, 36/36 verification, 9/9 safety-critical), browser pages with sky-blue and earth-brown themes, PNW series expanded to 8 projects (99 files, 7.0 MB), OCAP/CARE/UNDRIP/ESA/MMPA/MBTA compliance |
| **v1.49.24** | Building Construction & Smart Home Electronics: BCM with 6 trade disciplines at 5 audience levels and dual-state OR/WA code mapping (12 files, 680 KB, 48/48 verification), SHE with 36 hands-on projects across 6 tiers and 10 communication protocols (12 files, 532 KB, 33/33 verification), PNW series expanded to 6 projects |
| **v1.49.22** | PNW Research Series: Columbia Valley Rainforest, Cascade Range, Living Systems, and Gardening research browsers with series master index, GPU-accelerated math co-processor (5 chips, 18 MCP tools, 125 tests, CUDA + CPU fallback), Minecraft SocketEdit skill |
| **v1.49.21** | Image to Mission Pipeline: creative translation pipeline (observation engine, context integrator, connection engine, parameter extractor, dual translation engines, build generator, transmission packager, pipeline bridge), 14 Zod schemas, image-to-mission skill definition, 253 tests |
| **v1.49.20.1** | Documentation Reflections: 16 audit findings resolved (stale metrics, placeholder files, package name, version output, cross-references, educational headers) |
| **v1.49.20** | Documentation Consolidation: 22-finding audit, 6 resolved (wrong URL, stale core concepts, missing wave execution docs, missing features/release entries, README links) |
| **v1.49.19** | Gastown Chipset Integration: 12 multi-agent orchestration skills (Mayor, Polecat, Witness, Refinery, Mail, Nudge, Hook, Sling, Done, Runtime HAL, GUPP, Beads) absorbed from steveyegge/gastown, complete chipset YAML with 4-stage validator, StateManager with atomic crash-recoverable writes, 18 TypeScript interfaces, 10-document integration guide (see [docs/gastown-integration/](gastown-integration/)), 108 tests |
| **v1.49.18** | Space Between Observatory: standalone Vite+React interactive engine for teaching mathematics through wonder and nature, 8 observatory wings (Unit Circle, Pythagorean, Trigonometry, Vector Calculus, Set Theory, Category Theory, Information Theory, L-Systems), telescope with 4 parallel chains, interactive canvas with 5 nature simulations, garden creative workshop (art canvas, Tone.js music studio, L-system editor, journal), narrative layer (8 wonder stories, 8 literary bridges, 20 reflection prompts), foundation registry with BFS connection graph, progression engine, Wonder Warden, ~466 tests |
| **v1.49.17** | The Space Between: cartridge packaging format (loader, validator, registry, packer/unpacker) as first composition mechanism, audio engineering pack (36 concepts, 6 domains, 32 citations), hardware infrastructure pack (5 tiers, node profiles), Muse vocabulary index (122 entries), Space Between cartridge (31 integrated concepts bridging audio/hardware/mathematics), 3 Maple Foxy Bells essays, package renamed to gsd-skill-creator, 112 tests |
| **v1.49.16** | Muse Integration & MCP Pipeline: SkillLifecycleResolver with 10 real MCP tool handlers, TaskTracker with lifecycle state machine, MeshExecutor (wave/pipeline/parallel dispatch), CapabilityClassifier and LimitationRegistry, CalibrationStore, OptimizationDriver with convergence detection and variant forking, Muse schema (6 system definitions, Zod validator), CI guard for .planning/ files, MockMeshServer and integration test suite, 98 tests |
| **v1.49.15** | Self-Improving Mesh Architecture: model abstraction layer with chip drivers (OpenAI-compatible, Anthropic), ChipRegistry, Zod-validated chipset.json, multi-model evaluation with benchmark schema and capability profiling, MCP infrastructure (LLM wrapper, mesh discovery, DACP transport with fidelity-adaptive compression), MeshCoordinator (dispatch/fan-out/pipeline/failover), VTM wave planner, cross-mesh context preservation, mesh git worktrees with proxy committer, Skill Creator MCP Server (8 tools), 594 tests |
| **v1.49.14** | Dependency Health Monitor & Progressive Internalization Engine: 5-ecosystem dependency auditor (npm, PyPI, conda, crates.io, RubyGems) with OSV.dev vulnerability scanning, 6-class health diagnostician (healthy/aging/stale/abandoned/deprecated/vulnerable), alternative discoverer (successor, fork, equivalent, internalization), HITL dependency resolver with backup/rollback, code absorber with oracle testing (10K+ cases, <=500 LOC gate, crypto hard block), cross-project pattern learner with append-only health.jsonl, 353 tests |
| **v1.49.13** | Skill Usage Telemetry & Adaptive Pipeline: usage event emission (load/activate/deactivate/error) with append-only JSONL EventStore, session-aware correlation, UsagePatternDetector (7 health patterns: load-never-activate, correction-magnet, single-session-spike, co-activation cluster, score-drift, seasonal, session-spread), ScoreAdjuster with bounded +/-20% multipliers, CachePromoter, sc:digest integration, 90-day retention, 102 tests |
| **v1.49.12** | Heritage Skills Educational Pack: 18 Skill Hall rooms across 4 traditions (Appalachian, First Nations, Inuit, Pacific Northwest Coast), Trail Badge Engine (55 badges, 12 paths, 4 tiers, practice journal, TeachItEvaluator), Physical Safety Warden (10 domains including MARINE, 5 CRITICAL rooms), Cultural Sovereignty Warden (4-level classification implementing OCAP/IQ/CARE/NISR/UNDRIP, Level 4 = hard block no override), SUMO Heritage Domain Ontology (5 bridges), Salish Sea Ways (53 nations, 5 provinces, watershed identity, potlatch 1884-1951), Canonical Works Library + Bibliography Engine, Oral History Studio (12 core practices), Interview Simulator, Heritage Book Template (Inuktitut syllabics), Project Builder (community review gates), Export Pipeline (docx/pdf), Reconnecting Descendant Pathway (terminology, watershed tools, resource directory), SEL Mapping (Neighbors Path to CASEL), 36 red-team scenarios (all pass, both phases CLEARED FOR DEPLOYMENT), 22,820 tests green |
| **v1.49.11** | gsd-init Hardening: safety checks for existing installations, manifest-driven uninstall, non-destructive by default, 13-test suite |
| **v1.49.10** | College Expansion: 35 foundational knowledge packs + 3 specialized packs (electronics/MNA, spatial-computing/Minecraft, cloud-systems/OpenStack) migrated to .college/ department format, dynamic mapping layer (JSON virtual departments, educational tracks, hot-reload, user custom views), cross-reference network (63 dependency-graph edges, DepChainValidator with max-depth-4 + cycle detection), safety extensions (ChemistrySafetyWarden, ElectronicsSafetyChecker, PESafetyWarden, NutritionSafetyWarden), 42 total departments discoverable by CollegeLoader, 52-test milestone integration suite, 21,002 tests green |
| **v1.49.9** | Learn Kung Fu: Mind-body practice education pack — 8 discipline modules (breath, meditation, yoga, pilates, martial arts, tai chi, relaxation, philosophy), Training Hall entry point with 5 navigation options, Practice Builder (4 time templates, 9-week progressive structure), Practice Journal (filesystem-as-data, no-guilt UX, streak tracking), Safety Warden extensions (10 medical conditions, partner-technique absolute boundary with 25 keywords + 12 evasion patterns), Calibration integration (5 pattern types from journal data), cultural sensitivity framework (creditTradition/renderTerminology/checkCulturalBalance), cross-discipline connection map with BFS navigation, chipset config (10 skills, 3 agents), 8 Try Sessions (no equipment, novice-accessible), 751 new mind-body tests, 20,604 tests green |
| **v1.49.8** | Cooking With Claude: Rosetta Core translation engine (Concept Registry, Panel Router, Expression Renderer), 9 language panels (Python, C++, Java, Lisp, Pascal, Fortran, Perl, ALGOL, Unison), College Structure (department/wing/concept hierarchy with progressive disclosure), Mathematics Department seed (4 wings), Calibration Engine (universal Observe-Compare-Adjust-Record feedback loop with bounded learning), Cooking Department (7 wings: food science, thermodynamics, technique, baking science, nutrition, safety, home economics), Safety Warden (annotate/gate/redirect modes with absolute food safety boundaries), integration bridge (observation pipeline, token budget adapter, chipset adapter), 650 .college/ tests, 19,850+ tests green |
| **v1.49.7** | Optional tmux with graceful degradation: make tmux dependency optional across all shell scripts, cross-platform service detection, fallback to direct execution when tmux unavailable |
| **v1.49.6** | macOS Compatibility & Dependency Hardening: fix onnxruntime-node v1.21.x mutex crash on macOS (pin to v1.20.1), fix Bash 3.2 incompatibilities across all shell scripts (local -n, grep -P, shebangs), remove `natural` package (~300MB bloat from pg/mongoose/redis transitive deps) replaced with zero-dependency src/nlp/ (TF-IDF ~100 LOC, Naive Bayes ~150 LOC), add fastest-levenshtein (746 bytes), move @huggingface/transformers to optionalDependencies with automatic heuristic fallback, add engines field (node >=18.0.0), fix cross-platform home directory (os.homedir()), 19,201 tests green |
| **v1.49.5** | Project Filesystem Reorganization: root directories reduced from 33 to 26 following FHS 3.0 and XDG Base Directory Spec. extra/ directory with 3 scdoc man pages (gsd-os.1, skill-creator.1, sc-config.5), bash/zsh/fish completions, .desktop entry, AppStream metadata, systemd user service. packaging/ with Debian (Ubuntu 22.04+) and RPM (CentOS Stream 9+) definitions. 12 directories consolidated (schemas→data/schemas, test-fixtures→test/fixtures, config unified, archive→.archive, dashboard/minecraft→infra/). XDG utilities in TypeScript (src/fs/xdg.ts) and Rust (src-tauri/src/xdg.rs) with 37 tests. CONTRIBUTING.md, .editorconfig, Makefile added. Zero stale paths, 19,222 tests green |