@oussamadouhou/agent-enforcement 0.1.6 → 0.1.7

package/docs/EVIDENCE_SYSTEM.md

@@ -191,6 +191,49 @@ When you try to mark a todo complete without evidence:
 
 ---
 
+## Best Practices
+
+### Sequential Completion (Recommended)
+
+**Workflow:**
+1. Execute the task (run tool/command)
+2. Capture output in your response
+3. Provide complete evidence block
+4. Mark todo as `completed`
+5. Move to next task
+
+**Why sequential?**
+- Clearer evidence trail per task
+- Easier to verify each completion
+- Reduces batch-completion confusion
+- Helps with debugging if validation fails
+
+**Example:**
+```markdown
+I'll check the working directory.
+[tool: bash] pwd
+Output: /home/user/project
+
+**Evidence for todo-1**:
+**Execution**: `pwd`
+**Verification**: Path `/home/user/project` returned
+**Checklist**:
+- [x] Tool executed successfully
+- [x] Output captured
+- [x] Result matches expected
+- [x] No workarounds used
+- [x] Independently verifiable
+**Trust**: 🟢 HIGH | ✅ Ground Truth
+
+[tool: todowrite] mark todo-1 as completed
+
+Now moving to todo-2...
+```
+
+**Note:** Sequential completion is optional but recommended. The plugin validates evidence regardless of whether you complete tasks one-by-one or in batches.
+
+---
+
 ## Technical Details
 
 ### Hook Points

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@oussamadouhou/agent-enforcement",
-  "version": "0.1.6",
+  "version": "0.1.7",
   "description": "Evidence enforcement hooks for OpenCode and Claude Code",
   "main": "dist/plugin.js",
   "type": "module",