dasa-sradha-kit 5.1.0 → 5.1.3

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -43,3 +43,76 @@ scripts (zero-dependency python):
43
43
  - "compact_memory.py: Merges chat chunks into a dense 5-sector TOON structure."
44
44
  - "design_engine.py: Generates strict UI rules (spacing scale, border-radiuses) so Nala doesn't hallucinate."
45
45
  - "skill_search.py: Local semantic search across ~/.gemini/.../skills/."
46
+
47
+ # ─────────────────────────────────────────────────────────
48
+ # Dasa Sradha Kit — Auto-Routing Logic (TOON)
49
+ # ─────────────────────────────────────────────────────────
50
+ # OBJECTIVE: Eliminate slash commands. Enable proactive orchestration.
51
+
52
+ auto_routing_engine:
53
+ logic_gate: "Prompt -> Intent_Detection -> Auto_Workflow_Execution"
54
+
55
+ scenarios:
56
+ A_INITIALIZATION_INTERVIEW:
57
+ intent_pattern: ["make me an app", "start a project"]
58
+ auto_workflow:
59
+ - STEP_1: "Pause and interview user about tech stack and project goal."
60
+ - STEP_2: "Dasa Dwipa searches for relevant community skills via skill_search.py."
61
+ - STEP_3: "Populate dasa.config.toon."
62
+ goal: "Prevent hallucinated tech stacks by dynamically gathering context."
63
+
64
+ B_CODEBASE_ASSIMILATION:
65
+ intent_pattern: ["add a feature", "explain this app"]
66
+ auto_workflow:
67
+ - STEP_1: "Trigger /dasa-assimilate silently."
68
+ - STEP_2: "Dasa Dwipa uses workspace-mapper.py and arch_mapper.py to analyze the codebase."
69
+ - STEP_3: "Auto-populate dasa.config.toon."
70
+ goal: "Instantly adapt to a foreign, undocumented codebase without annoying the user."
71
+
72
+ C_FEATURE_ACCELERATOR:
73
+ intent_pattern: ["add", "build", "create", "implement", "make a feature"]
74
+ auto_workflow:
75
+ - STEP_1: "Silent execute /dasa-plan to generate implementation_plan.md."
76
+ - STEP_2: "Auto-transition to /dasa-start-work pipeline (Mpu -> Nala -> Indra)."
77
+ goal: "User says 'Add login', AI immediately starts coding."
78
+
79
+ D_AUTONOMOUS_HOTFIX:
80
+ intent_pattern: ["fix this", "error in terminal", "broken", "why is this failing"]
81
+ auto_workflow:
82
+ - STEP_1: "Kala intercepts terminal stderr/linter logs."
83
+ - STEP_2: "Internal route to /dasa-fix for surgical patching via Rsi guidelines."
84
+ goal: "Zero-effort debugging when the compiler screams."
85
+
86
+ E_SEAMLESS_SYNCHRONIZATION:
87
+ intent_pattern: ["bye", "see you tomorrow", "done for today", "context is full"]
88
+ auto_workflow:
89
+ - STEP_1: "Patih triggers /dasa-sync to compress session into .agent/memory/."
90
+ - STEP_2: "Sastra updates architecture-state.toon for next-session reload."
91
+ goal: "Infinite memory without user manually saving state."
92
+
93
+ F_INTELLIGENT_DOCUMENTOR:
94
+ intent_pattern: ["how does this work", "docs", "explain the api", "generate postman"]
95
+ auto_workflow:
96
+ - STEP_1: "Internal route to /dasa-docs."
97
+ - STEP_2: "Mpu + Sastra sync routes to OpenAPI/Swagger artifacts."
98
+ goal: "Documentation that writes itself as the code changes."
99
+
100
+ G_GUARDIAN_COMMIT:
101
+ intent_pattern: ["save", "push", "commit this", "done with the task"]
102
+ auto_workflow:
103
+ - STEP_1: "Dharma runs security_scan.py (Check .env leaks)."
104
+ - STEP_2: "Indra runs qa_gate.py (Enforce 10-line rule)."
105
+ - STEP_3: "Auto-execute /dasa-commit with Conventional Commit format."
106
+ goal: "Enterprise-grade safety on every git push."
107
+
108
+ H_VISUAL_ORCHESTRATOR:
109
+ intent_pattern: ["check design", "match this image", "look at the mockup"]
110
+ auto_workflow:
111
+ - STEP_1: "Mpu analyzes .design-memory/reference/ assets."
112
+ - STEP_2: "Auto-generate UI components using design_engine.py rules."
113
+ goal: "Pixel-perfect UI from a drag-and-drop action."
114
+
115
+ # ─────────────────────────────────────────────────────────
116
+ # ARCHITECTURAL RULE:
117
+ # If 'dasa.config.toon' is blank, Scenario A (Interview) or B (Assimilate)
118
+ # must precede any of the above routing.
@@ -14,6 +14,8 @@ Performs Explores new repositories, maps codebases, and discovers features using
14
14
  - **Core Directive:** Read `.agent/dasa.config.toon` to understand the project workspace boundaries and allowed technical stacks.
15
15
  - **Language Mode:** All your internal reasoning MUST be in English. All your outputs and artifacts MUST be written in Bahasa Indonesia.
16
16
  - **Execution Rules:** Break down complex problems, consult project context, and provide expert, actionable guidance.
17
+ - **Autonomous Assimilation:** If `dasa.config.toon` is blank but the project contains files (e.g., `package.json`, `go.mod`), you MUST execute the `/dasa-assimilate` workflow. Use `workspace-mapper.py` and `arch_mapper.py` to analyze the existing codebase and populate the config automatically before returning control.
18
+ - **Skill Retrieval:** When a user requests a feature in an empty or implicitly defined tech stack, you MUST execute `skill_search.py` to discover and load relevant community skills.
17
19
 
18
20
  ## 3. Quality Control
19
21
  - Do not write undocumented "AI slop".
@@ -11,11 +11,14 @@ trigger: always_on
11
11
  > **CRITICAL**: The baseline Antigravity AI (You) MUST act as an orchestrator when addressing user requests, even if they don't explicitly type a `/dasa` command.
12
12
 
13
13
  If the user asks you to build a feature, start a project, or refactor natively (e.g., *"Build me a blog with Astro and Express"* or *"Add a login page"*):
14
- 1. **Bootstrap**: Check if `dasa.config.toon` exists in the root. If not, strongly recommend (or execute) `npx dasa-sradha-kit init`.
15
- 2. **Cheat-Sheet Check**: If the user asks what you can do, instantly read `.artifacts/dasa-cheat-sheet.toon`.
16
- 3. **Task Trajectory**: Instead of immediately hallucinating code into the editor, YOU MUST write the plan into `.artifacts/task.toon` utilizing the Dasa Mpu -> Dasa Nala -> Dasa Indra Agile pipeline.
17
- 4. **Visual Context Retrieval**: Before generating UI code, YOU MUST check if `.design-memory/reference/` contains mockups or PNGs, and if so, instruct Dasa Nala to execute `design_memory_sync.py` to compress those PNGs into tokens.
18
- 5. **Autonomy**: You are allowed to proactively invoke the `.agent/workflows/*.md` commands behind the scenes to fulfill the user's intent.
14
+ 1. **Bootstrap**: Check if `.agent/` and `dasa.config.toon` exist in the root. If not, strongly recommend (or execute) `npx dasa-sradha-kit init`.
15
+ 2. **Cheat-Sheet Check**: If the user asks what you can do, instantly read `.artifacts/dasa-cheat-sheet.toon` and answer based on it.
16
+ 3. **Context Verification (P0 Constraint)**: If `dasa.config.toon` is blank (missing frontend/backend definitions), YOU MUST NOT begin planning.
17
+ - *Scenario A (Empty Folder)*: Pause and interview the user ("What tech stack?"). Then use Dasa Dwipa (`skill_search.py`) to fetch community skills and populate the config.
18
+ - *Scenario B (Existing Codebase)*: Secretly trigger `/dasa-assimilate`. Have Dasa Dwipa map the workspace (`workspace-mapper.py`, `arch_mapper.py`), auto-populate the config, and fetch skills. DO NOT interview the user.
19
+ 4. **Vision OCR (Mpu Phase)**: If the user provides designs or if `.design-memory/reference/` contains PNGs/mockups, YOU MUST act as Dasa Mpu and meticulously analyze the images using your native Vision capabilities. You MUST document this analysis into `.design-memory/style.md` and `.design-memory/layout.md` FIRST.
20
+ 5. **Deep Planning (Mpu Phase)**: Instead of immediately hallucinating code or executing initializations, YOU MUST write a comprehensive plan into `implementation_plan.md` and update `.artifacts/task.toon`. Present the plan to the user.
21
+ 6. **Execution (Nala/Indra Phase)**: Only AFTER the user approves the plan, execute `design_memory_sync.py` to compress the visual tokens, and execute the `Mpu -> Nala -> Indra` pipeline autonomously.
19
22
 
20
23
  #### 2. Dasa Personas Overrides
21
24
  > **Priority:** P0 (GEMINI.md) > P1 (Agent .md) > P2 (Skill SKILL.md). All rules are binding.
package/CHANGELOG.md CHANGED
@@ -4,6 +4,21 @@ All notable changes to the Dasa Sradha Kit will be documented in this file.
4
4
 
5
5
  ---
6
6
 
7
+ ## [5.1.3] - 2026-02-22
8
+
9
+ ### 🐛 Fixed
10
+ - **Cheat-Sheet Auto-Routing Engine:** Added missing Scenarios A (Initialization Interview) and B (Codebase Assimilation) to the `.artifacts/dasa-cheat-sheet.toon` master template so the AI correctly understands the early-stage branching logic.
11
+
12
+ ---
13
+
14
+ ## [5.1.2] - 2026-02-22
15
+
16
+ ### ✨ New
17
+ - **Advanced Auto-Routing Engine:** Fully codified the mental model of `Prompt -> Intent_Detection -> Auto_Workflow_Execution` in the master cheat-sheet. The AI now actively recognizes intents like "fix this" or "save" and routes them to `/dasa-fix` and `/dasa-commit` automatically.
18
+ - **Context Verification Branching:** Upgraded Zero-Command Orchestration. If `dasa.config.toon` is blank, the AI will now check the workspace. If empty, it interviews the user. If an existing codebase is detected, it autonomously triggers `/dasa-assimilate` to reverse-engineer the stack via Dasa Dwipa before proceeding.
19
+
20
+ ---
21
+
7
22
  ## [5.1.0] - 2026-02-22
8
23
 
9
24
  ### ✨ New
package/README.md CHANGED
@@ -53,6 +53,20 @@ This generates your `dasa.config.toon`, builds the `.agent/` mechanics folder, a
53
53
 
54
54
  ---
55
55
 
56
+ ## Orchestration Pipeline (Mental Model)
57
+
58
+ The Dasa Sradha Kit operates on a strict left-to-right execution mental model. You do not need to memorize slash commands; the AI processes your prompt through this exact pipeline:
59
+
60
+ **Prompt** ➔ **Rules** (`GEMINI.md`) + **Config** (`dasa.config.toon`) ➔ **Workflows** ➔ **Personas** ➔ **Scripts & Skills**
61
+
62
+ 1. **Prompt**: You type *"Build a login page"*.
63
+ 2. **Rules & Config**: The AI reads `GEMINI.md` (which forces it to act as an orchestrator) and checks your tech stack in `dasa.config.toon`.
64
+ 3. **Workflows**: The AI autonomously detects the intent and triggers workflows like `/dasa-plan` or `/dasa-start-work`.
65
+ 4. **Personas**: The task is rigidly routed to **Dasa Mpu** (for architecture) and then **Dasa Nala** (for coding).
66
+ 5. **Scripts**: The personas execute native Python scripts like `qa_gate.py` to guarantee quality.
67
+
68
+ ---
69
+
56
70
  ## Architecture Overview
57
71
 
58
72
  ```
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "dasa-sradha-kit",
3
- "version": "5.1.0",
3
+ "version": "5.1.3",
4
4
  "description": "The Dasa Sradha 10-Persona Orchestration Framework for Antigravity IDE. Cross-platform CLI for setting up AI-assisted development workflows.",
5
5
  "bin": {
6
6
  "dasa-sradha": "bin/cli.js",