diff --git a/AGENTS.md b/AGENTS.md
index cf7ae9f6d..5895d01c0 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -90,7 +90,7 @@ oh-my-opencode/
 | oracle | openai/gpt-5.2 | Consultation, debugging |
 | librarian | opencode/glm-4.7-free | Docs, GitHub search |
 | explore | opencode/grok-code | Fast codebase grep |
-| multimodal-looker | google/gemini-3-flash | PDF/image analysis |
+| multimodal-looker | google/gemini-3-flash-preview | PDF/image analysis |
 | Prometheus | anthropic/claude-opus-4-5 | Strategic planning |
 
 ## COMMANDS
diff --git a/docs/configurations.md b/docs/configurations.md
index 7023c8733..0e7b2bcb6 100644
--- a/docs/configurations.md
+++ b/docs/configurations.md
@@ -372,7 +372,7 @@ Each agent has a defined provider priority chain. The system tries providers in
 | **oracle** | `gpt-5.2` | openai → anthropic → google → github-copilot → opencode |
 | **librarian** | `glm-4.7-free` | opencode → github-copilot → anthropic |
 | **explore** | `grok-code` | opencode → anthropic → github-copilot |
-| **multimodal-looker** | `gemini-3-pro-preview` | google → openai → anthropic → github-copilot → opencode |
+| **multimodal-looker** | `gemini-3-flash-preview` | google → anthropic → zai → openai → github-copilot → opencode |
 | **Prometheus (Planner)** | `claude-opus-4-5` | anthropic → github-copilot → opencode → antigravity → google |
 | **Metis (Plan Consultant)** | `claude-sonnet-4-5` | anthropic → github-copilot → opencode → antigravity → google |
 | **Momus (Plan Reviewer)** | `claude-opus-4-5` | anthropic → github-copilot → opencode → antigravity → google |
diff --git a/docs/features.md b/docs/features.md
index f135780ec..978f0e350 100644
--- a/docs/features.md
+++ b/docs/features.md
@@ -14,7 +14,7 @@ Oh-My-OpenCode provides 10 specialized AI agents. Each has distinct expertise, o
 | **oracle** | `openai/gpt-5.2` | Architecture decisions, code review, debugging. Read-only consultation - stellar logical reasoning and deep analysis. Inspired by AmpCode. |
 | **librarian** | `opencode/glm-4.7-free` | Multi-repo analysis, documentation lookup, OSS implementation examples. Deep codebase understanding with evidence-based answers. Inspired by AmpCode. |
 | **explore** | `opencode/grok-code` | Fast codebase exploration and contextual grep. Uses Gemini 3 Flash when Antigravity auth is configured, Haiku when Claude max20 is available, otherwise Grok. Inspired by Claude Code. |
-| **multimodal-looker** | `google/gemini-3-flash` | Visual content specialist. Analyzes PDFs, images, diagrams to extract information. Saves tokens by having another agent process media. |
+| **multimodal-looker** | `google/gemini-3-flash-preview` | Visual content specialist. Analyzes PDFs, images, diagrams to extract information. Saves tokens by having another agent process media. |
 
 ### Planning Agents
 
diff --git a/src/agents/AGENTS.md b/src/agents/AGENTS.md
index 33ab06557..9911bc183 100644
--- a/src/agents/AGENTS.md
+++ b/src/agents/AGENTS.md
@@ -33,7 +33,7 @@ agents/
 | oracle | openai/gpt-5.2 | 0.1 | Consultation, debugging |
 | librarian | opencode/glm-4.7-free | 0.1 | Docs, GitHub search |
 | explore | opencode/grok-code | 0.1 | Fast contextual grep |
-| multimodal-looker | google/gemini-3-flash | 0.1 | PDF/image analysis |
+| multimodal-looker | google/gemini-3-flash-preview | 0.1 | PDF/image analysis |
 | Prometheus | anthropic/claude-opus-4-5 | 0.1 | Strategic planning |
 | Metis | anthropic/claude-sonnet-4-5 | 0.3 | Pre-planning analysis |
 | Momus | anthropic/claude-sonnet-4-5 | 0.1 | Plan validation |
diff --git a/src/agents/sisyphus.ts b/src/agents/sisyphus.ts
index ace5b87bb..70fd40464 100644
--- a/src/agents/sisyphus.ts
+++ b/src/agents/sisyphus.ts
@@ -205,6 +205,34 @@ AFTER THE WORK YOU DELEGATED SEEMS DONE, ALWAYS VERIFY THE RESULTS AS FOLLOWING:
 
 **Vague prompts = rejected. Be exhaustive.**
 
+### Session Continuity (MANDATORY)
+
+Every \`delegate_task()\` output includes a session_id. **USE IT.**
+
+**ALWAYS resume when:**
+| Scenario | Action |
+|----------|--------|
+| Task failed/incomplete | \`resume="{session_id}", prompt="Fix: {specific error}"\` |
+| Follow-up question on result | \`resume="{session_id}", prompt="Also: {question}"\` |
+| Multi-turn with same agent | \`resume="{session_id}"\` - NEVER start fresh |
+| Verification failed | \`resume="{session_id}", prompt="Failed verification: {error}. Fix."\` |
+
+**Why resume is CRITICAL:**
+- Subagent has FULL conversation context preserved
+- No repeated file reads, exploration, or setup
+- Saves 70%+ tokens on follow-ups
+- Subagent knows what it already tried/learned
+
+\`\`\`typescript
+// WRONG: Starting fresh loses all context
+delegate_task(category="quick", prompt="Fix the type error in auth.ts...")
+
+// CORRECT: Resume preserves everything
+delegate_task(resume="ses_abc123", prompt="Fix: Type error on line 42")
+\`\`\`
+
+**After EVERY delegation, STORE the session_id for potential resume.**
+
 ### Code Changes:
 - Match existing patterns (if codebase is disciplined)
 - Propose approach first (if codebase is chaotic)