open-agents-ai 0.186.6 → 0.186.7

package/README.md

@@ -1072,6 +1072,67 @@ The identity kernel maintains a persistent self-model across sessions, the refle
 | L8 | Darwin Gödel Machine: Open-Ended Self-Improvement (2025) | [arxiv:2505.22954](https://arxiv.org/abs/2505.22954) |
 | L8 | i-MENTOR: Intrinsic Motivation Exploration (2025) | [arxiv:2505.17621](https://arxiv.org/abs/2505.17621) |
 
+## Agent Immune System — Constraint Enforcement & Pressure Resistance
+
+<div align="right"><a href="#top">back to top</a></div>
+
+Open Agents includes a behavioral immune system that prevents the agent from making pattern-matched mistakes under pressure. Inspired by biological immune systems: constraints are the antibodies, pressure detection is the inflammatory response, and memory injection is the recall mechanism.
+
+### Constraint Enforcement (`.oa/constraints.json`)
+
+Machine-readable rules checked **before every tool execution**:
+
+```json
+{
+  "constraints": [
+    {
+      "id": "no-reward-hack",
+      "trigger": "file_write|file_edit",
+      "pattern": "NEVER say|ALWAYS say",
+      "target_files": ["prompts/**/*.md"],
+      "action": "warn",
+      "message": "This looks like a reward-hacking directive. Fix the architecture, not the prompt."
+    }
+  ]
+}
+```
+
+| Action | Behavior |
+|--------|----------|
+| `block` | Prevents tool execution entirely, returns error to model |
+| `warn` | Executes tool but emits warning in agent's next turn context |
+| `log` | Silent recording to audit log, no interruption |
+
+Constraints are scoped: global (`~/.open-agents/constraints.json`), project (`.oa/constraints.json`), or session (ephemeral).
+
+### Pressure-Aware Decision Gate
+
+When the user is frustrated (detected via keyword matching), a brief `<reflection>` cue is injected into the agent's system prompt for ONE turn:
+
+```
+<reflection>The user is very frustrated. Pause. Check your constraints
+and past feedback before writing code. The fastest fix is often the wrong fix.</reflection>
+```
+
+This is NOT a block — it's a speed bump that prompts deliberation when the agent is most likely to cut corners. Zero overhead when no pressure is detected.
+
+| Pressure Level | Detection | Response |
+|---------------|-----------|----------|
+| **none** | Normal messages | No cue (zero tokens) |
+| **moderate** | Frustration signals | "Verify your change addresses the root cause" |
+| **high** | Strong frustration + urgency | "Pause. Check constraints before acting" |
+
+### How It Works Together
+
+```
+User (frustrated): "fix this broken shit"
+  → Pressure gate detects "high" → injects reflection cue
+  → Model proposes file_edit on prompts/system.md with "NEVER say..."
+  → Constraint checker matches "no-reward-hack" → emits warning
+  → Model sees warning on next turn → reconsiders approach
+  → Model fixes the architecture instead of adding a prompt hack
+```
+
 ## Context Compaction — Research-Backed Memory Management
 
 <div align="right"><a href="#top">back to top</a></div>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "open-agents-ai",
-  "version": "0.186.6",
+  "version": "0.186.7",
   "description": "AI coding agent powered by open-source models (Ollama/vLLM) — interactive TUI with agentic tool-calling loop",
   "type": "module",
   "main": "./dist/index.js",