@cubis/foundry 0.3.18 → 0.3.19

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/antigravity/rules/GEMINI.md

@@ -13,6 +13,18 @@ This file defines mandatory behavior for Antigravity projects installed via `cbx
 - Skills: `.agent/skills`
 - Rules file: `.agent/rules/GEMINI.md`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `.agent/rules/GEMINI.md`, plus any additional rule files loaded).
+2. Workflow decision (`/workflow` or direct mode) and why it was chosen.
+3. Agent routing decision (`@agent` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Workflow-First Contract
 
 1. If the user explicitly requests a slash command, run that workflow first.

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/codex/rules/AGENTS.md

@@ -9,6 +9,18 @@ This file defines mandatory behavior for Codex projects installed via `cbx workf
 - Skills: `.agents/skills`
 - Rules file: `AGENTS.md`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `AGENTS.md`, plus any additional rule files loaded).
+2. Workflow decision (`$workflow-*` or direct mode) and why it was chosen.
+3. Agent routing decision (`$agent-*` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Skill-Based Workflow
 
 1. Codex operates primarily through **Skills** and **Rules**.

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/copilot/rules/AGENTS.md

@@ -9,6 +9,18 @@ This file defines mandatory behavior for GitHub Copilot projects installed via `
 - Skills: `.github/skills`
 - Rules file (project): `.github/copilot-instructions.md`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `.github/copilot-instructions.md`, plus any additional rule files loaded).
+2. Workflow decision (`/workflow` or direct mode) and why it was chosen.
+3. Agent routing decision (`@agent` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Workflow-First Contract
 
 1. If the user explicitly requests a slash command, run that workflow first.

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/copilot/rules/copilot-instructions.md

@@ -9,6 +9,18 @@ This file defines mandatory behavior for GitHub Copilot projects installed via `
 - Skills: `.github/skills`
 - Rules file (project): `.github/copilot-instructions.md`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `.github/copilot-instructions.md`, plus any additional rule files loaded).
+2. Workflow decision (`/workflow` or direct mode) and why it was chosen.
+3. Agent routing decision (`@agent` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Workflow-First Contract
 
 1. If the user explicitly requests a slash command, run that workflow first.

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/cursor/rules/.cursorrules

@@ -9,6 +9,18 @@ This file defines mandatory behavior for Cursor projects installed via `cbx work
 - Skills: `.cursor/skills`
 - Rules file: `.cursorrules`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `.cursorrules`, plus any additional rule files loaded).
+2. Workflow decision (`/workflow` or direct mode) and why it was chosen.
+3. Agent routing decision (`@agent` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Smart Skill Selection (TIER 0)
 
 Before starting ANY task, the agent MUST:

package/Ai Agent Workflow/workflows/agent-environment-setup/platforms/windsurf/rules/.windsurfrules

@@ -9,6 +9,18 @@ This file defines mandatory behavior for Windsurf projects installed via `cbx wo
 - Skills: `.windsurf/skills`
 - Rules file: `.windsurfrules`
 
+## Startup Transparency (Required)
+
+Before executing workflows, agents, or code edits, publish a short `Decision Log` that is visible to the user:
+
+1. Rule file(s) read at startup (at minimum `.windsurfrules`, plus any additional rule files loaded).
+2. Workflow decision (`/workflow` or direct mode) and why it was chosen.
+3. Agent routing decision (`@agent` or direct mode) and why it was chosen.
+4. Skill loading decision (skill names loaded) and why they were chosen.
+
+If routing changes during the task, publish a `Decision Update` before continuing.
+Keep this user-visible summary concise and factual; do not expose private chain-of-thought.
+
 ## 2) Smart Skill Selection (TIER 0)
 
 Before starting ANY task, the agent MUST:

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cubis/foundry",
-  "version": "0.3.18",
+  "version": "0.3.19",
   "description": "Cubis Foundry CLI for workflow-first AI agent environments",
   "type": "module",
   "bin": {