PARA Workspace Logo # PARA Workspace **The Workspace Framework for Humans & AI Agents** [![License](https://img.shields.io/badge/license-MIT-green.svg)](https://opensource.org/licenses/MIT) [![Version](https://img.shields.io/badge/version-1.9.2-blue.svg)](./CHANGELOG.md) ![Type](https://img.shields.io/badge/type-workspace_framework-blueviolet.svg) [![Antigravity](https://img.shields.io/badge/Antigravity-verified-E37400?logo=google&logoColor=white)](https://antigravity.google/) πŸ‡ΊπŸ‡Έ English β€’ πŸ‡»πŸ‡³ TiαΊΏng Việt β€’ πŸ‡¨πŸ‡³ δΈ­ζ–‡ β€’ πŸ‡ͺπŸ‡Έ EspaΓ±ol β€’ πŸ‡«πŸ‡· FranΓ§ais
--- | Section | Description | | :--- | :--- | | [🌌 Overview](#-overview) | What it is, three foundational principles | | [πŸ“‚ Architecture](#-architecture) | Repo structure + generated workspace structure | | [πŸ“₯ Installation](#-installation) | Prerequisites, setup, profiles, troubleshooting | | [🧠 The Kernel](#-the-kernel) | Invariants, heuristics, contracts | | [πŸ› οΈ CLI Reference](#️-cli-reference) | All CLI commands | | [πŸ“‘ Workflow Catalog](#-workflow-catalog) | 32 governed workflows | | [πŸ›‘οΈ Rule Catalog](#️-rule-catalog) | 14 governance rules | | [🧩 Skill Catalog](#-skill-catalog) | 21 reusable skills | | [πŸ”Œ Tool System](#-tool-system-v180) | Install external agentic plugins | | [🧩 Task Management](#-task-management-hybrid-3-file-model) | Hybrid 3-File model | | [πŸ”„ Upgrading](#-upgrading-versions) | Auto update + clean slate | | [πŸ—ΊοΈ Roadmap](#️-roadmap) | Version history + planned features | ## 🌌 Overview **PARA Workspace** is an open-source workspace framework that defines how humans and AI agents organize knowledge and collaborate on projects. It ships as a **repo** containing a kernel (constitution), CLI tools, and templates β€” which generates **workspaces** where you actually work. The kernel enforces invariants and heuristics so every workspace is predictable, auditable, and agent-friendly. ### Three Foundational Principles 1. **Repo β‰  Workspace** β€” The repo contains governance (kernel, CLI, templates). It never contains user data. 2. **Workspace = Runtime** β€” Generated by `para init`, each workspace is a standalone instance where you and your agent work. 3. **Kernel = Constitution** β€” Immutable rules that all workspaces follow. Changes require an RFC + version bump. ```mermaid flowchart TD R["πŸ›οΈ Repo\n(Constitution + Compiler)"] W["πŸ’» Workspace\n(OS Runtime)"] A["πŸ€– Agent\n(Execution Environment)"] R -->|para init| W W -->|agent attach| A style R fill:#4a90d9,stroke:#2c5f8a,color:#fff style W fill:#50c878,stroke:#2e8b57,color:#fff style A fill:#ff8c42,stroke:#cc6633,color:#fff ``` --- ## πŸ“‚ Architecture ### Repo Structure (This Repository) ``` para-workspace/ β”œβ”€β”€ .github/ # πŸ€– CI/CD β€” validate-pr.yml, CODEOWNERS β”œβ”€β”€ rfcs/ # πŸ“ RFC Process β€” TEMPLATE.md, status in header β”œβ”€β”€ kernel/ # 🧠 Constitution β”‚ β”œβ”€β”€ KERNEL.md β”‚ β”œβ”€β”€ invariants.md # 10 hard rules (MAJOR bump) β”‚ β”œβ”€β”€ heuristics.md # 10 soft conventions β”‚ β”œβ”€β”€ schema/ # workspace, project, backlog, catalog schemas β”‚ └── examples/ # valid/ + invalid/ compliance vectors β”œβ”€β”€ cli/ # πŸ”§ Compiler β”‚ β”œβ”€β”€ para # Entry point (Bash 3.2+ compatible) β”‚ β”œβ”€β”€ lib/ # logger.sh, validator.sh, rollback.sh, fs.sh β”‚ └── commands/ # init, scaffold, status, migrate, archive, install, update β”œβ”€β”€ templates/ # πŸ“¦ Scaffolding & Governed Libraries β”‚ β”œβ”€β”€ common/agents/ # Centralized workflows/, rules/, skills/ + catalog.yml β”‚ β”‚ └── projects/ # .project.yml template β”‚ └── profiles/ # dev, general presets β”œβ”€β”€ tests/ # πŸ§ͺ kernel/ + cli/ integration tests β”œβ”€β”€ docs/ # πŸ“– Documentation β”‚ β”œβ”€β”€ architecture/ # Architecture: overview, kernel, ecosystem β”‚ β”œβ”€β”€ guides/ # How-to: development, planning, meta-project β”‚ β”œβ”€β”€ reference/ # Lookup: CLI, workflows, project-rules β”‚ β”œβ”€β”€ rules/ # Individual rule documentation (11 files) β”‚ └── workflows/ # Individual workflow docs (22 files) β”œβ”€β”€ CONTRIBUTING.md β”œβ”€β”€ VERSIONING.md β”œβ”€β”€ CHANGELOG.md └── VERSION ``` ### Workspace Structure (Generated by `para init`) ``` / β”œβ”€β”€ Projects/ # Goal-oriented tasks β”‚ β”œβ”€β”€ my-app/ # Standard project (type: standard) β”‚ β”‚ β”œβ”€β”€ repo/ # Source code (git repo) β”‚ β”‚ β”œβ”€β”€ artifacts/ # Plans, tasks, decisions β”‚ β”‚ β”œβ”€β”€ sessions/ # Session logs β”‚ β”‚ β”œβ”€β”€ docs/ # Project documentation β”‚ β”‚ └── project.md # Project contract β”‚ └── my-ecosystem/ # Ecosystem project (type: ecosystem) β€” v1.6.0+ β”‚ β”œβ”€β”€ artifacts/ # Cross-project plans & backlog β”‚ β”œβ”€β”€ sessions/ # Session logs β”‚ β”œβ”€β”€ docs/ # Strategy & shared docs β”‚ └── project.md # satellites: [app, api, ...], NO repo/ β”œβ”€β”€ Areas/ # Ongoing responsibilities (e.g., health, finance) β”‚ β”œβ”€β”€ Workspace/ # Master session log, audits, SYNC queue β”‚ └── Learning/ # Shared knowledge (from /learn) β”œβ”€β”€ Resources/ # Reference material & tools β”‚ β”œβ”€β”€ ai-agents/ # Kernel snapshot + governed library snapshots β”‚ └── references/ # The original PARA repository (read-only) β”œβ”€β”€ Archive/ # Cold storage for completed items β”œβ”€β”€ _inbox/ # Temporary landing zone for external downloads β”œβ”€β”€ .agents/ # Governed library copies (Auto-synced) β”‚ β”œβ”€β”€ rules.md # Workspace Rules Trigger Index (always loaded) β”‚ β”œβ”€β”€ skills.md # Workspace Skills Trigger Index (v1.6.2+) β”‚ β”œβ”€β”€ rules/ # Active agent rules (.md) β”‚ β”œβ”€β”€ skills/ # Active agent skills (.md, /scripts) β”‚ └── workflows/ # Active agent workflows (.md) β”œβ”€β”€ .para/ # System state (DO NOT EDIT) β”‚ β”œβ”€β”€ archive/ # Smart Archive vault for deprecated files β”‚ β”œβ”€β”€ migrations/ # Migration execution logs β”‚ β”œβ”€β”€ backups/ # Date-stamped workflow & project backups β”‚ └── audit.log # Action history β”œβ”€β”€ para # Bootstrapper CLI └── .para-workspace.yml # Workspace root metadata config ``` --- ## πŸ“₯ Installation ### Prerequisites - **An AI Agent platform** (see compatibility table below) - **Git** (required β€” to clone and update) - **Bash** 3.2+ (Linux/macOS native, Windows via Git Bash or WSL) ### Platform Compatibility | Platform | Integration Point | Status | Notes | | :----------------- | :----------------------------- | :---------- | :--------------------------------------------------- | | Google Antigravity | `.agents/` (skills, workflows) | βœ… Verified | Designed and tested for this platform | | Claude Code | CLAUDE.md + `.agents/` | βšͺ Untested | May read `.agents/rules/` directly β€” needs validation | | Cursor | `.cursor/rules/` | βšͺ Untested | Theory: copy rules to `.cursor/rules/` | | VS Code + Copilot | `.github/copilot-instructions` | βšͺ Untested | Theory: instructions only, no auto-trigger | ### Step 1: Clone & Install **Bash (Linux / macOS / Windows Git Bash / WSL):** ```bash # Clone repo into the correct location mkdir -p Resources/references git clone https://github.com/pageel/para-workspace.git Resources/references/para-workspace # Set executable permissions chmod +x Resources/references/para-workspace/cli/para chmod +x Resources/references/para-workspace/cli/commands/*.sh # Initialize your workspace with a profile ./Resources/references/para-workspace/cli/para init --profile=dev --lang=en ``` **PowerShell (alternative for Windows):** ```powershell mkdir -Force Resources\references git clone https://github.com/pageel/para-workspace.git Resources\references\para-workspace # Then open Git Bash or WSL at workspace root: ./Resources/references/para-workspace/cli/para init --profile=dev --lang=en ``` ### Step 2: Verify ```bash ./para status # βœ… If you see a health report β†’ installation successful ``` > **What just happened?** > > 1. The repo lives at `Resources/references/para-workspace/` β€” it's a read-only reference source, not a user project. > 2. `chmod +x` ensures all CLI scripts are executable (required on Linux/macOS). > 3. `para init` creates the full PARA directory structure, runs `install.sh` automatically to sync kernel, workflows, governance rules, and generates a `./para` wrapper. > 4. You can now use `./para` from your workspace root for all commands. ### Updating ```bash # Pull latest from GitHub and re-sync workspace ./para update # Preview changes before applying ./para update --dry-run ``` This will `git pull` the repo, run version-gated migrations, and re-sync all governed libraries. Existing user-customized files are backed up to `.bak` before overwriting (Smart Sync). If the install fails mid-operation, all changes are automatically rolled back. **What happens during update:** 1. `git pull` fetches latest code (self-restarts if scripts changed) 2. `migrate.sh` runs version-gated migration steps (only what's needed) 3. `install.sh` syncs kernel, workflows, rules, skills to workspace (with atomic rollback) 4. Audit log is updated in `.para/audit.log` ### Troubleshooting | Problem | Solution | | :------------------------------- | :-------------------------------------------------------- | | **macOS: permission denied** | Run `chmod +x` on CLI scripts (Step 3 above) | | **Windows: file lock on update** | See [Windows Recovery](#windows-recovery) below | | **Stale workspace (v1.3.x)** | Use [Manual Clean Slate](#mechanism-2-manual-clean-slate) | #### Windows Recovery If `para update` fails on Windows due to NTFS file locking: ```cmd cd Resources\references\para-workspace git checkout -- . git pull origin main cd ..\..\.. .\para install ``` ### Available Profiles | Profile | Description | Best For | | :---------------------------------------------------- | :---------------------------------- | :---------------------- | | [`general`](./templates/profiles/general/README.md) | Minimal PARA structure | Personal PKM | | [`dev`](./templates/profiles/dev/README.md) | Technical Areas + AI tooling | Software developers | ### What `para init` Does - βœ… Creates `Projects/`, `Areas/`, `Resources/`, `Archive/`, and `_inbox/` - βœ… Sets **executable permissions** on all CLI scripts - βœ… Runs **`install.sh`** automatically, which: - Syncs **governed workflows** from `catalog.yml` into `.agents/workflows/` - Syncs **governed rules** from `catalog.yml` into `.agents/rules/` - Syncs **workspace rules index** into `.agents/rules.md` (trigger index) - Syncs **skills** into `.agents/skills/` (if profile includes them) - Creates **`./para`** wrapper at workspace root - Backs up conflicting files to `.bak` - βœ… Generates **`.para-workspace.yml`** with kernel version tracking - βœ… Initializes **`.para/`** state (audit.log, migrations/, backups/) for full auditability --- ## 🧠 The Kernel The Kernel is the **constitution** of PARA Workspace β€” the rules that all workspaces must follow. ### Invariants (Hard Rules β€” change = MAJOR bump) | # | Rule | | --- | ------------------------------------------------------------------- | | I1 | PARA directory structure is mandatory | | I2 | Hybrid 3-file task model (backlog = canonical, hot lane, /end sync) | | I3 | kebab-case project naming | | I4 | No active tasks = inactive project | | I5 | Areas contain no runtime tasks | | I6 | Archive is immutable cold storage | | I7 | Seeds are raw ideas, not tasks | | I8 | No loose files at workspace root | | I9 | Resources are read-only references | | I10 | Repo ↔ Workspace separation | | I11 | Workflow language compliance | ### Heuristics (Soft Rules β€” change = MINOR/PATCH) | # | Guideline | | --- | ---------------------------------------------------------- | | H1 | Naming conventions (kebab-case, PascalCase) | | H2 | Context loading priority | | H3 | Semantic versioning with approval levels | | H4 | Standard project directory structure | | H5 | Beads lifecycle | | H6 | VCS & Git boundaries | | H7 | Cross-project references + Ecosystem + Meta-Project (v1.7.6) | | H8 | Workflow kernel compatibility | | H9 | Governed libraries require `catalog.yml` with `kernel_min` | | H10 | Knowledge Items β€” schema, scope, slug validation (v1.7.0) | ### Kernel ↔ Workspace Contracts | File | Schema | Enforced By | | ---------------------------- | ----------------------- | ------------------------------ | | `.para-workspace.yml` | `workspace.schema.json` | `para init`, `para status` | | `Projects/*/.project.yml` | `project.schema.json` | `para scaffold`, `para review` | | `artifacts/tasks/backlog.md` | `backlog.schema.json` | `para verify` | | `*/catalog.yml` | `catalog.schema.json` | `para install`, `para update` | --- ## πŸ› οΈ CLI Reference ```bash # Core Commands para init [--profile] [--lang] # Create workspace para status [--json] # System health para update # Auto-update & migrate para scaffold # Create structured paths para install [--force] # Sync governed libraries para archive # Graduation review para migrate [--from] [--to] # Workspace migration # Configuration para config [key] [value] # Manage workspace settings # Tool Management (v1.8.0) para install-tool # Install a plugin from registry para remove-tool # Remove installed plugin para list-tools # List installed plugins # MCP Server Management (v1.8.2) para mcp-setup # Auto-configure MCP in IDE para mcp-list # List MCP-capable tools para mcp-remove # Remove MCP config from IDE # Agent Capabilities @[/para-workflow] list # Manage workflows @[/para-rule] list # Manage rules ``` **Platform Support:** Linux, macOS (Bash 3.2+), Windows (Git Bash / WSL). --- ## πŸ“‘ Workflow Catalog | Command | Description | | :-------------------------------------------------------- | :--------------------------------------------------------- | | **[`/backlog`](./docs/workflows/backlog.md)** | Manage project tasks via canonical backlog.md | | **[`/backup`](./docs/workflows/backup.md)** | Backup workflows, rules, config, and project data | | **[`/brainstorm`](./docs/workflows/brainstorm.md)** | Collaborative ideation with structured output | | **[`/para-config`](./docs/workflows/para-config.md)** | Manage workspace configuration | | **[`/end`](./docs/workflows/end.md)** | Close session with PARA classification + automated cleanup | | **[`/inbox`](./docs/workflows/inbox.md)** | Categorize files from `_inbox/` into PARA | | **[`/install`](./docs/workflows/install.md)** | Intelligent installer (handles updates/merges) | | **[`/learn`](./docs/workflows/learn.md)** | Capture lessons into Areas/Learning | | **[`/merge`](./docs/workflows/merge.md)** | Semantic merge for workflow conflicts | | **[`/new-project`](./docs/workflows/new-project.md)** | Initialize new project with scaffolding | | **[`/open`](./docs/workflows/open.md)** | Start session with context + plan phase loading | | **[`/para`](./docs/workflows/para.md)** | Master controller for workspace management | | **[`/docs`](./docs/workflows/docs.md)** | Generate, review, and publish technical documentation | | **[`/para-audit`](./docs/workflows/para-audit.md)** | Macro Assessor for tracking workspace structural drift | | **[`/para-rule`](./docs/workflows/para-rule.md)** | Manage, install, and standardize agent rules | | **[`/para-workflow`](./docs/workflows/para-workflow.md)** | Manage, install, and standardize agent workflows | | **[`/plan`](./docs/workflows/plan.md)** | Create, review, and update implementation plans | | **[`/push`](./docs/workflows/push.md)** | Fast commit and push to GitHub | | **[`/qa`](./docs/workflows/qa.md)** | Systematic Red Team Q&A review loop for plans and specs | | **[`/release`](./docs/workflows/release.md)** | Pre-release quality gate | | **[`/retro`](./docs/workflows/retro.md)** | Project retrospective before archiving | | **[`/spec`](./docs/workflows/spec.md)** | Write structured specifications before coding | | **[`/update`](./docs/workflows/update.md)** | Agent-guided safe workspace update with error recovery | | **[`/verify`](./docs/workflows/verify.md)** | Verify task completion with walkthroughs | | **[`/para-knowledge`](./docs/workflows/para-knowledge.md)** | Manage Knowledge Items β€” dashboard, create, audit, archive | | **[`/para-skill`](./docs/workflows/para-skill.md)** | Co-Author engine for creating, validating, and testing AI Agent skills | | **[`/write`](./docs/workflows/write.md)** | Write deep-dive content with structured sidecar templates | | **[`/logs`](./docs/workflows/logs.md)** | Session telemetry diagnostics and context budget analytics | | **[`/staging`](./docs/workflows/staging.md)** | Stage workspace assets into a project's repo (v1.8.11) | | **[`/vibecode`](./docs/workflows/vibecode.md)** | Manage interactive/loop execution modes for coding tasks (v1.8.11) | | **[`/scan-sec`](./docs/workflows/scan-sec.md)** | Scan codebase for security vulnerabilities (v1.8.11) | | **[`/resource`](./docs/workflows/resource.md)** | Orchestrate resource import, graph generation, and pattern learning (v1.8.11) | --- ## πŸ›‘οΈ Rule Catalog Rules govern agent behavior, security, and compliance. Loaded on-demand via a Two-Tier trigger index (~200 tokens). Skills are loaded via a parallel skills trigger index (v1.6.2). See [Rules & Context Loading](./docs/reference/project-rules.md) for the loading mechanism. | Rule | Description | Priority | | :-------------------------------------------------------------------------------- | :------------------------------------------------------ | :----------- | | **[`governance`](./docs/rules/governance.md)** | Kernel invariants, scope containment, safety guardrails | πŸ”΄ Critical | | **[`vcs`](./docs/rules/vcs.md)** | Git safety: commit, branch, merge, PR, secrets | πŸ”΄ Critical | | **[`knowledge`](./docs/rules/knowledge.md)** | KI operations β€” write gate, approval, namespace (v1.7.0)| πŸ”΄ Critical | | **[`graph-first-policy`](./docs/rules/graph-first-policy.md)** | Proactive code analysis using para-graph MCP tools | πŸ”΄ Critical | | **[`hybrid-3-file-integrity`](./docs/rules/hybrid-3-file-integrity.md)** | 6 constraints (C1–C6) for task management | 🟑 Important | | **[`agent-behavior`](./docs/rules/agent-behavior.md)** | Proactive Trigger Check, Context Recovery (v1.6.2) | 🟑 Important | | **[`context-rules`](./docs/rules/context-rules.md)** | Agent Index Loading (rules + skills), Two-Tier trigger | 🟑 Important | | **[`para-discipline`](./docs/rules/para-discipline.md)** | PARA architecture compliance, 7 rules | 🟑 Important | | **[`agent-persona`](./docs/rules/agent-persona.md)** | Custom conversational style and agent personality | 🟑 Important | | **[`artifact-standard`](./docs/rules/artifact-standard.md)** | Plans, walkthroughs, persistent mirroring | 🟒 Standard | | **[`naming`](./docs/rules/naming.md)** | kebab-case, PascalCase, camelCase conventions | 🟒 Standard | | **[`versioning`](./docs/rules/versioning.md)** | SemVer, autonomy levels, multi-location sync | 🟒 Standard | | **[`exports-data`](./docs/rules/exports-data.md)** | Data export: `_inbox/`, UTF-8 BOM, naming | 🟒 Standard | | **[`tool-routing`](./docs/rules/tool-routing.md)** | Heuristic guide for selecting Native API, Bash, and MCP | 🟒 Standard | --- ## 🧩 Skill Catalog Skills are reusable knowledge modules loaded on-demand via the skills trigger index. Unlike rules (which enforce constraints), skills provide **templates, patterns, and reference material**. | Skill | Description | | :--- | :--- | | **[PARA Kit](./docs/skills/para-kit.md)** | PARA workspace structure reference β€” schema, layout, kernel governance, intelligence routing | | **[Formatting](./docs/skills/formatting.md)** | Tables, diagrams, tree listings, ASCII box art | | **[Page Map](./docs/skills/page-map.md)** | Website visual structure management using PAGE_MAP.md and BLUEPRINT.md | | **[PARA Skill](./docs/skills/para-skill.md)** | Governance skill for creating and validating PARA skills via Co-Author engine | | **[Plan Templates](./docs/skills/plan.md)** | Detail Plan & Roadmap templates (Sidecar, v1.7.8) | | **[Docs Templates](./docs/skills/docs.md)** | Architecture, CLI, Strategy templates (Sidecar, v1.7.8) | | **[Brainstorm Templates](./docs/skills/brainstorm.md)** | Decision & Research templates (Sidecar, v1.7.12) | | **[Write Templates](./docs/skills/write.md)** | Content type templates and writing rules (Sidecar, v1.7.14) | | **[Harness Guards](./docs/skills/harness.md)** | Guard catalog and auto-scan protocol for context-aware safety guards (Sidecar, v1.7.15) | | **[Spec Templates](./docs/skills/spec.md)** | Spec templates, assumption surfacing, boundary definitions (Sidecar, v1.7.16) | | **[QA Review Templates](./docs/skills/qa.md)** | Red Team personas, dimension checklists, and QA report templates (Sidecar, v1.8.6) | | **[TDD Guidelines](./docs/skills/tdd.md)** | Strict Test-Driven Development workflow rules and anti-patterns (v1.8.7) | | **[Logs Audit Extensions](./docs/skills/logs.md)** | Audit templates for TDD/Spec compliance verification (Sidecar, v1.8.7) | | **[HTML Renderer](./docs/skills/html-renderer.md)** | Modular HTML rendering engine for themed static docs & graph views (v1.8.9) | | **[New Project](./docs/skills/new-project.md)** | Sidecar skill for the /new-project workflow (Sidecar, v1.8.10) | | **[para-graph](./docs/skills/para-graph.md)** | Centralized Graph Intelligence Router and semantic memory curation (Sidecar, v1.8.10) | | **[Staging Templates](./docs/skills/staging.md)** | Sidecar data for the /staging template release (Sidecar, v1.8.11) | | **[Vibecode Execution Templates](./docs/skills/vibecode.md)** | Sidecar data for the /vibecode execution (Sidecar, v1.8.11) | | **[Vulnerability Scanner Templates](./docs/skills/scan-sec.md)** | Sidecar data for the /scan-sec scanner (Sidecar, v1.8.11) | | **[Resource Study Templates](./docs/skills/resource.md)** | Sidecar data for the /resource learning (Sidecar, v1.8.11) | | **[Sidecar Skill Governance](./docs/skills/sidecar-skill.md)** | Architectural rules and standards for sidecar skills (v1.8.11) | --- ## πŸ—οΈ Rule Architecture β€” Two-Tier Loading & Defense-in-Depth Rules and skills aren't dumped into context all at once. PARA Workspace uses a **progressive disclosure** architecture that loads indices on-demand, saving ~90% of token budget while maintaining full compliance. ### How It Works: Two-Tier Trigger Index (v1.6.2) ```mermaid flowchart TD START["πŸš€ /open starts session"] --> R subgraph T1["Tier 1 β€” ALWAYS"] R["πŸ“œ .agents/rules.md\n~20 lines Β· ~200 tokens"] S["🧩 .agents/skills.md\n~10 lines Β· ~100 tokens"] end R --> MEMO1["Agent memorizes triggers\nβ†’ loads rules ON DEMAND"] S --> MEMO2["Agent memorizes triggers\nβ†’ loads skills ON DEMAND"] START --> S MEMO1 --> CHECK MEMO2 --> CHECK subgraph T2["Tier 2 β€” CONDITIONAL"] CHECK{"Check project.md\nagent map"} CHECK -->|agent.rules: true| PR["πŸ“œ Project rules.md"] CHECK -->|agent.skills: true| PS["🧩 Project skills.md"] CHECK -->|neither| SKIP["Skip β€” no project indices"] end style T1 fill:#1a1a2e,stroke:#4a90d9,color:#e0e0e0 style T2 fill:#1a1a2e,stroke:#50c878,color:#e0e0e0 style START fill:#ff8c42,stroke:#cc6633,color:#fff style CHECK fill:#4a90d9,stroke:#2c5f8a,color:#fff ``` > πŸ’‘ **Total cost: ~350 tokens** (vs ~2,000 if all loaded upfront) ### Defense-in-Depth: 4-Layer Protection After a long conversation, AI agents lose context (truncation). PARA Workspace has **4 independent layers** to ensure rules and skills survive: ```mermaid block-beta columns 4 block:header:4 columns 4 h1["Layer"] h2["What"] h3["Where"] h4["Survives"] end l1["1"] w1["Index instructions"] p1["agent-behavior.md Β§4"] s1["⚠️ Lost"] l2["2"] w2["Safety block"] p2["/open Step 8 report"] s2["βœ… Checkpoint"] l3["3"] w3["Pre-flight"] p3["Step 0: re-read"] s3["βœ… Disk"] l4["4"] w4["File guards"] p4[" inline"] s4["βœ… In-file"] style s1 fill:#d32f2f,color:#fff style s2 fill:#388e3c,color:#fff style s3 fill:#388e3c,color:#fff style s4 fill:#388e3c,color:#fff ``` Layer 3 uses **Proactive Trigger Check** (v1.6.2): BEFORE any side-effect, agent scans all trigger tables and loads matching rules/skills FIRST. Layer 4 supports **6 guard types**: `STATUS GATE`, `MANDATORY`, `HARNESS GUARD`, `FILE GUARD` (C1-C3, I9, GOVERNED), `WORKFLOW GATE`, `CONTEXT RECOVERY`. **Scenario: Agent forgets rules after truncation** ```mermaid sequenceDiagram participant C as πŸ’¬ Long Conversation participant T as βœ‚οΈ Truncation participant P as πŸš€ Agent runs /push participant D as πŸ“ Agent edits done.md C->>T: Context window full T--xP: ⚠️ Rules lost! T--xD: ⚠️ Rules lost! Note over P: Layer 3 catches P->>P: Step 0 re-reads .agents/rules.md P->>P: VCS rules loaded β†’ safe commit βœ… Note over D: Layer 4 catches D->>D: Reads D->>D: Obeys C2 β†’ append only βœ… ``` > πŸ“– Full architecture: [Context Recovery](./docs/architecture/context-recovery.md) Β· [Defense-in-Depth](./docs/architecture/defense-in-depth.md) Β· [Rule Layers](./docs/architecture/rule-layers.md) --- ## 🧩 Task Management (Hybrid 3-File Model) PARA Workspace uses a proprietary **Hybrid 3-File Architecture** (v1.5.3: Hot Lane) to solve the AI Agent "Context Window vs. Amnesia" problem. Instead of forcing the agent to read one massive backlog file every time it opens a project (which wastes tokens and causes hallucination), tasks are distributed across three highly specialized files: ``` artifacts/tasks/ β”œβ”€β”€ backlog.md # πŸ“Œ CANONICAL β€” Operational Authority. All strategic tasks live here. β”œβ”€β”€ sprint-current.md # πŸ”₯ HOT LANE β€” Agent-writable buffer for ad-hoc quick tasks. └── done.md # βœ… ARCHIVE β€” Append-only log with origin tags (#backlog / #session). ``` ### How It Works 1. **Backlog Summary** β€” `/open` reads only the Summary table (~10 lines) + top active items via `grep`. Never the full file. 2. **Hot Lane** β€” Agent writes ad-hoc quick tasks (`- [ ] Fix CSS`) to `sprint-current.md` and ticks `[x]` when done. Strategic tasks from backlog are read directly, not copied. 3. **`/end` is the Sole Sync Point** β€” At session close, `/end` runs Hot Lane Sync: - Quick `[x]` tasks β†’ `done.md` (tagged `#session`) - Quick `[ ]` tasks β†’ ask user: keep or promote to backlog? - **Smart Suggest** β†’ reads session log, finds mentioned task IDs, suggests marking strategic tasks as Done (tagged `#backlog`) 4. **Zero Ceremony During Coding** β€” No `/backlog update` needed mid-session. Just code. 5. **`/backlog clean` = Compress, Not Delete** β€” Done items are compressed into 1 line per plan in backlog's `βœ… Completed` section. Details live in `done.md` (grouped by plan, linking to `plans/done/`). Lookup chain: `backlog β†’ done.md β†’ plans/done/`. --- ## πŸ“š Knowledge System (v1.7.0) AI agents lose context between sessions. **Knowledge Items (KIs)** are a core feature of [Google Antigravity](https://antigravity.google/docs/knowledge) β€” the AI coding agent platform β€” that provides persistent memory across sessions, projects, and conversations. KIs live **outside** the workspace, in Antigravity's platform-managed KI Store (`~/.gemini/antigravity/knowledge/`). This mirrors PARA Workspace's foundational principle: just as the **repo governs but does not contain** user data, PARA Workspace **governs KI operations** (schema, rules, lifecycle workflows) without owning the storage layer. ```mermaid flowchart LR subgraph Platform["Antigravity Platform"] KI["πŸ“¦ KI Store\n~/.gemini/antigravity/\nknowledge/"] end subgraph WS["PARA Workspace"] GOV["πŸ›‘οΈ Governance\nschema Β· rules\n/para-knowledge"] end GOV -->|governs| KI style Platform fill:#1a1a2e,stroke:#ff8c42,color:#e0e0e0 style WS fill:#1a1a2e,stroke:#4a90d9,color:#e0e0e0 style KI fill:#ff8c42,stroke:#cc6633,color:#fff style GOV fill:#4a90d9,stroke:#2c5f8a,color:#fff ``` ### How It Works ```mermaid flowchart LR S1["πŸ” Session 1\nAgent discovers insight"] S2["♻️ Session 2\nAuto-loaded"] SN["⚑ Session N\nApplies instantly"] KI[("πŸ’Ύ KI Store\nPlatform-managed")] S1 -->|save as KI| KI KI -->|auto-inject| S2 S2 -->|refine| KI KI -->|auto-inject| SN style S1 fill:#4a90d9,stroke:#2c5f8a,color:#fff style S2 fill:#50c878,stroke:#2e8b57,color:#fff style SN fill:#ff8c42,stroke:#cc6633,color:#fff style KI fill:#9b59b6,stroke:#7d3c98,color:#fff ``` Unlike session logs (lost after truncation) or learning files (project-scoped), KIs are: - **Auto-injected** at session start by the platform - **Cross-project** β€” workspace-scoped KIs apply to all projects - **Graph-ready** β€” domain Γ— purpose taxonomy with `relates_to` and `concepts` fields ### KI Taxonomy Each KI is classified along two axes β€” **domain** (what) Γ— **purpose** (how): ``` β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β” β”‚ PURPOSE β”‚ β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”¬β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€ β”‚ context β”‚ reference β”‚ pitfall β”‚ playbook β”‚ β”‚ (priming) β”‚ (lookup) β”‚ (gotchas) β”‚ (how-to) β”‚ β”Œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”Όβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€ β”‚ workspace β”‚ PARA arch β”‚ kernel specs β”‚ update bugs β”‚ audit proc β”‚ β”‚ engineeringβ”‚ tech stack β”‚ API docs β”‚ dep gotchas β”‚ deploy SOP β”‚ D β”‚ operations β”‚ infra setup β”‚ CI/CD ref β”‚ env traps β”‚ rollback β”‚ O β”‚ content β”‚ writing std β”‚ style guide β”‚ SEO traps β”‚ publish β”‚ M β”‚ strategy β”‚ org context β”‚ roadmap β”‚ scope creep β”‚ planning β”‚ A β”‚ ... β”‚ β”‚ β”‚ β”‚ β”‚ I β”œβ”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜ β”‚ β”‚ β”‚ β”‚ N β”‚ (open β€” define your own) β”‚ β”‚ β”‚ β”‚ β””β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”΄β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”€β”˜ ``` > **Domain** is an open string β€” define custom domains as needed. > **Purpose** is a fixed enum: `context`, `reference`, `pitfall`, `playbook`. ### Lifecycle & Integration | Workflow | KI Integration | |:---------|:---------------| | `/open` | Platform-injected KI context, scope matching for report | | `/plan` | Pitfall KIs β†’ Risks, Playbook KIs β†’ Phase references | | `/end` | Suggests creating/updating KIs from session insights | | `/brainstorm` | Option F: Extract insight as KI | | `/retro` | Graduates cross-project patterns to KIs | | `/para-knowledge` | Full lifecycle: create, update, audit, archive | ### Governance KI operations are governed by **7 rules (KR1–KR7)** and validated against **H10** (11 clauses): - **KR1 Write Gate**: Only `/para-knowledge` workflow can create/modify KIs - **KR2 User Approval**: Every KI creation requires explicit user consent - **KR3 Namespace**: `para_` prefix reserved for system KIs - **KR5 Recoverability**: All KI operations are reversible - **KR7 Ephemeral Ban**: No ephemeral file paths in KI references (v1.7.5) > πŸ“– Full schema: [`ki.schema.json`](./kernel/schema/ki.schema.json) Β· Governance: [`heuristics.md` H10](./kernel/heuristics.md) --- ## πŸ”Œ Tool System (v1.8.0) PARA Workspace supports an extensible **Dynamic Tool System** that allows you to install external, language-agnostic plugins (like `para-graph`) directly into your workspace. Tools are managed via a central registry (`registry/tools.yml`) and are installed as standalone tarballs. ### How it works 1. **Zero Global Dependencies**: Tools are installed locally into `.para/tools/` for isolation. 2. **Multi-Runtime Support**: The CLI auto-generates wrapper scripts (e.g. `repo/cli/commands/graph.sh`) that know how to invoke Node, Python, or binary executables. 3. **Dev/Prod Fallback**: If a tool's source code exists inside the workspace (Dev mode), the wrapper routes execution there. Otherwise, it falls back to the extracted tarball (Prod mode). ### Available Tools | Tool | Description | | :--- | :--- | | **[`para-graph`](https://github.com/pageel/para-graph)** | Hybrid Code-Knowledge Graph for PARA Workspace | > πŸ’‘ Version is resolved dynamically from `registry/tools.yml` β€” always installs the latest registered release. ### Usage ```bash # Install the para-graph plugin (structural code analysis & MCP server) ./para install-tool para-graph # List installed tools ./para list-tools # Run the installed tool ./para graph --help # Remove tool ./para remove-tool para-graph ``` ### Tool Intelligence Installer (v1.8.1) Tools can bundle AI intelligence (workflows, skills, rules) directly in their `tool.manifest.yml`: ```yaml agents: workflows: - source: templates/agents/workflows/para-graph.md target: para-graph.md version: "1.8.0" skills: - source: templates/agents/skills/graph-enrichment/ target: graph-enrichment/ version: "1.0.0" ``` When you run `./para install-tool `, the CLI will automatically parse this manifest and prompt you to install the bundled intelligence. You can use `--agents` to install only the agents, or `--no-agents` to skip the prompt. `remove-tool` will also offer to clean up any bundled agents it installed. ### MCP Auto-Setup (v1.8.2) Tools can declare an MCP server in their manifest via the `mcp:` block. PARA Workspace provides three CLI commands to manage MCP configurations across IDEs: ```bash # Auto-configure MCP server for a tool in your IDE ./para mcp-setup # Print the JSON snippet without modifying IDE config ./para mcp-setup --print-only # List all tools with MCP servers available ./para mcp-list # Remove an MCP server entry from IDE config ./para mcp-remove ``` **Supported IDEs:** | IDE | Config Path | | :--- | :--- | | Google Antigravity | `~/.gemini/antigravity/mcp_config.json` | | Claude Desktop | `~/.config/claude/claude_desktop_config.json` (Linux) | | Cursor | `~/.cursor/mcp.json` | > The setup uses **2-Tier path resolution** (Dev path β†’ Prod path) to point to the correct entry, and `jq`-based atomic JSON merge with automatic backup. --- ## πŸ”„ Upgrading Versions > πŸ“– **Note:** For a detailed breakdown of all features, fixes, and updates in each version, please read our [CHANGELOG](./CHANGELOG.md). PARA Workspace offers two official mechanisms to upgrade to a newer version: ### Mechanism 1: Auto Update (Recommended) For most healthy workspaces, the built-in update mechanism handles everything safely. ```bash ./para update ``` This will automatically pull the newest core code, run version-gated migrations (only steps relevant to your current version), sync the libraries, and securely move obsolete structural files to the `.para/archive/` vault without deleting your custom data (Smart Archive). User-customized rules are protected with `.bak` backups. ### Mechanism 2: Manual Clean Slate (For heavily modified workspaces) If your workspace is very old (v1.3.x) or has been heavily customized, start fresh: 1. Run `para init --profile=dev --lang=en` in a completely new directory. 2. Copy your old `Projects/` directories into the `_inbox/` of the new workspace. 3. Use the `/inbox` workflow or let your AI agent sort them into the new correct structure. --- ## πŸ—ΊοΈ Roadmap - [x] Cross-Project Sync Queue _(shipped in v1.3.6)_ - [x] Kernel Extraction & Repo Restructure _(shipped in v1.4.0)_ - [x] Governed Libraries, RFC Process, Workspace Runtime Safety _(shipped in v1.4.1)_ - [x] Landing Page `paraworkspace.dev` _(shipped in v1.4.1)_ - [x] Plan-Aware Workflows & Token Optimization _(shipped in v1.4.2)_ - [x] Automated Plan Cleanup via `/end [done]` _(shipped in v1.4.3)_ - [x] Safety Guardrails & Terminal Allowlist _(shipped in v1.4.5)_ - [x] Smart Archive & Version Migration _(shipped in v1.4.6)_ - [x] macOS Compatibility & Safe Migration Pipeline _(shipped in v1.4.7)_ - [x] Atomic Rollback, Dry-run Pipeline & README Rewrite _(shipped in v1.4.8)_ - [x] Centralized Backup & Workspace Cleanup _(shipped in v1.4.9)_ - [x] Project Rules Loading & Safe Update Workflow _(shipped in v1.5.0)_ - [x] Hybrid 3-File Synchronization & Fast Mode _(shipped in v1.5.1)_ - [x] Hybrid 3-File Integrity, Working Checkmarks & Docs Overhaul _(shipped in v1.5.2)_ - [x] Hot Lane Refactor, /end Sync Point & Token Optimization _(shipped in v1.5.3)_ - [x] Context Recovery, Workflow Pre-flight & Defense-in-Depth _(shipped in v1.5.4)_ - [x] Meta-Project & Ecosystem Support _(shipped in v1.6.0)_ - [x] Unified Strategy β†’ Plan Flow _(shipped in v1.6.1)_ - [x] Unified Agent Index β€” Skills Loading & Proactive Trigger Check _(shipped in v1.6.2)_ - [x] Central Gate β€” project.md as single source for context loading _(shipped in v1.6.3)_ - [x] Para-Kit Skill v1.1.0, Recursive Sync & Git Hash Detection _(shipped in v1.6.4)_ - [x] Update Flow Fix β€” Version Direction Detection & Migration History _(shipped in v1.6.5)_ - [x] **Knowledge System β€” KI schema, /knowledge workflow, graph-ready taxonomy** _(shipped in v1.7.0)_ - [x] **System KI Governed Lifecycle β€” namespace guard, template sync, CLI hooks** _(shipped in v1.7.1)_ - [x] **KI Index Schema Upgrade, Workflow Simplification & Knowledge Graph Seed** _(shipped in v1.7.2)_ - [x] **Agent Path Convention Fix (BUG-28) & Rule Frontmatter** _(shipped in v1.7.3)_ - [x] **Repo Path Standardization, Pending TODO Fix & Project Profile Skill** _(shipped in v1.7.4)_ - [x] **KR7 Ephemeral Ban & /knowledge β†’ /para-knowledge Rename** _(shipped in v1.7.5)_ - [x] **Skill Catalog Architecture & Meta-Project Type** _(shipped in v1.7.6)_ - [x] **Brainstorm Consolidation, Sidecar Skill & /docs** _(shipped in v1.7.7-1.7.8)_ - [x] **Proactive Trigger Check & Cognitive Bypass Fix** _(shipped in v1.7.9-1.7.10)_ - [x] **Backlog Governance & Extract Paradigm** _(shipped in v1.7.11-1.7.12)_ - [x] **VERSIONS.yml Migration & Anti-Token Decay** _(shipped in v1.7.13)_ - [x] **Content Authoring Ecosystem & Session Telemetry** _(shipped in v1.7.14)_ - [x] **Harness Skill, Plan Status Gate & Roadmap Prefix Convention** _(shipped in v1.7.15)_ - [x] **Spec Workflow, Dual-Format Guards & CLI Fixes** _(shipped in v1.7.16)_ - [x] **Dynamic Tool System & para-graph integration** _(shipped in v1.8.0)_ - [x] **Tool Intelligence Installer & manifest-declared AI intelligence** _(shipped in v1.8.1)_ - [x] **MCP Auto-Setup β€” `mcp-setup`, `mcp-list`, `mcp-remove` CLI commands** _(shipped in v1.8.2)_ - [x] **QA Governance Workflow & Sidecar Skill Integration** _(shipped in v1.8.6)_ - [x] **TDD Governance Workflow** _(shipped in v1.8.7)_ - [x] **Ecosystem-wide Para-Graph MCP Integration β€” Memory Ops lifecycle, Project Sidecar Skill, 43% workflow coverage** _(shipped in v1.8.9)_ - [x] **Release Hardening, HTML Renderer Integration, Spec Rules & TDD Gate** _(shipped in v1.8.10)_ - [x] **Release Hardening: dirty-check template override prompts in install-tool CLI, conflict diff detection, and staging/scan-sec/vibecode/resource workflows catalog integration** _(shipped in v1.8.11)_ - [x] **Antigravity IDE 2.x and v1.x Config Migration & TDD hardening** _(shipped in v1.8.12)_ - [x] **Node Path Resolution (NVM/fnm/Volta) & Rule/KI Artifact Authority improvements** _(shipped in v1.8.13)_ - [x] **MCP Windows Path Conversion (BUG-37) & Plan Auto-Activation** _(shipped in v1.8.14)_ - [x] **Vibecode DSP & Template Consolidation** _(shipped in v1.8.15)_ - [ ] Department System _(v1.9.0 β€” planned)_ - [ ] Community & Trust Boundary _(v1.10.0 β€” planned)_ --- ## 🀝 Contributing See [CONTRIBUTING.md](./CONTRIBUTING.md) for guidelines. Key points: - Kernel invariant changes require an **RFC + MAJOR bump** β†’ See [rfcs/TEMPLATE.md](./rfcs/TEMPLATE.md) - Heuristic changes require a **PR + MINOR/PATCH bump** - All changes must pass test vectors in `kernel/examples/` --- ## πŸ“„ License & References This project is licensed under the MIT License. The `/scan-sec` security audit module (governed under [scan-sec SKILL.md](./templates/common/agents/skills/scan-sec/SKILL.md)) is built upon and references: - The [vbsec repository](https://github.com/tanviet12/vbsec) (core execution architecture). - The [OWASP Top 10 project](https://owasp.org/www-project-top-ten/) (security standards and classifications). ### Third-Party Dependencies #### Core CLI Utilities - [jq](https://jqlang.github.io/jq/) (JSON CLI parser, required for configuration updates). - [Git](https://git-scm.com/) (Required for version control and workspace updates). - [Node.js](https://nodejs.org/) (Runtime for graph analysis and HTML document renderers). #### Frontend CDN Libraries (used in HTML rendering) - [marked](https://github.com/markedjs/marked) (Markdown parser). - [mermaid](https://github.com/mermaid-js/mermaid) (Flowchart/diagram engine). - [force-graph](https://github.com/vasturiano/force-graph) (2D force-directed graph engine for code graph visualization). - [lucide](https://github.com/lucide-icons/lucide) (UI icons). - [Google Fonts](https://fonts.google.com/) (Inter, Outfit, Roboto, Fira Code typography). --- Built with ❀️ by **Pageel**. Standardizing the future of Agentic PKM. _Version: 1.9.2_