aebrer · m-aebrer · May 15, 2026 · May 14, 2026 · May 14, 2026 · May 14, 2026
diff --git a/package-lock.json b/package-lock.json
diff --git a/package.json b/package.json
@@ -34,7 +34,7 @@
 	"engines": {
 		"node": ">=20.0.0"
 	},
-	"version": "2.18.0",
+	"version": "2.19.0",
 	"dependencies": {
 		"@mariozechner/jiti": "^2.6.5",
 		"@dreb/coding-agent": "*",

diff --git a/packages/agent/package.json b/packages/agent/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@dreb/agent-core",
-	"version": "2.18.0",
+	"version": "2.19.0",
 	"description": "General-purpose agent with transport abstraction, state management, and attachment support",
 	"type": "module",
 	"main": "./dist/index.js",

diff --git a/packages/ai/package.json b/packages/ai/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@dreb/ai",
-	"version": "2.18.0",
+	"version": "2.19.0",
 	"description": "Unified LLM API with automatic model discovery and provider configuration",
 	"type": "module",
 	"main": "./dist/index.js",

diff --git a/packages/ai/test/image-tool-result.test.ts b/packages/ai/test/image-tool-result.test.ts
@@ -201,9 +201,10 @@ async function handleToolWithTextAndImageResult<TApi extends Api>(
 		const lowerContent = textContent.text.toLowerCase();
 		// Should mention details from the text (diameter/pixels)
 		expect(lowerContent.match(/diameter|100|pixel/)).toBeTruthy();
-		// Should also mention the visual properties (red and circle)
+		// Should also mention the visual color. Some models describe the object as a
+		// generic shape/figure instead of repeating "circle", but red is only present
+		// in the image and verifies that the visual block was read.
 		expect(lowerContent).toContain("red");
-		expect(lowerContent).toContain("circle");
 	}
 }
 

diff --git a/packages/coding-agent/README.md b/packages/coding-agent/README.md
@@ -346,6 +346,8 @@ The `subagent` tool delegates tasks to independent child agent processes. Each s
 
 **Agent definitions** live in `~/.dreb/agents/` (global) and `.dreb/agents/` (project). Each is a markdown file with YAML frontmatter specifying `name`, `model` (with provider fallback list), and optional `systemPrompt`. Built-in agents include `Explore` (read-only codebase exploration), `Sandbox` (restricted to `/tmp`), `feature-dev` (strong-tier coding), and several review agents.
 
+**Model availability probes:** When an agent definition specifies a fallback list (comma-separated models), each model is verified with a lightweight 1-token API call before the subagent is spawned. Models that fail the probe (rate limit, quota exhaustion, auth failure, timeout) are skipped with a loud log line, and the next fallback is tried. If all configured models fail, the parent session's model is used as a last resort. Per-invocation model overrides and single-model configs skip probing entirely.
+
 **Session metadata:** Each child process records its agent type in the session JSONL header (`agentType` field), providing an audit trail of which agent definition executed the work.
 
 ---

diff --git a/packages/coding-agent/package.json b/packages/coding-agent/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "@dreb/coding-agent",
-	"version": "2.18.0",
+	"version": "2.19.0",
 	"description": "Coding agent CLI with read, bash, edit, write tools and session management",
 	"type": "module",
 	"drebConfig": {
-Original file line number
+Diff line change
@@ Expand Up @@
     **Agent definitions** live in `~/.dreb/agents/` (global) and `.dreb/agents/` (project). Each is a markdown file with YAML frontmatter specifying `name`, `model` (with provider fallback list), and optional `systemPrompt`. Built-in agents include `Explore` (read-only codebase exploration), `Sandbox` (restricted to `/tmp`), `feature-dev` (strong-tier coding), and several review agents.
+    **Model availability probes:** When an agent definition specifies a fallback list (comma-separated models), each model is verified with a lightweight 1-token API call before the subagent is spawned. Models that fail the probe (rate limit, quota exhaustion, auth failure, timeout) are skipped with a loud log line, and the next fallback is tried. If all configured models fail, the parent session's model is used as a last resort. Per-invocation model overrides and single-model configs skip probing entirely.
     **Session metadata:** Each child process records its agent type in the session JSONL header (`agentType` field), providing an audit trail of which agent definition executed the work.
     ---
@@ Expand Down @@