@tdsoft-tech/aikit 0.1.9 → 0.1.11

package/dist/index.js

@@ -1766,6 +1766,95 @@ Before marking complete:
 2. Display summary to user
 3. Propose next actions
 4. Continue from where left off`
+  },
+  {
+    name: "one-shot",
+    description: "End-to-end autonomous task execution (beta)",
+    category: "core",
+    usage: "/one-shot <task description>",
+    examples: ["/one-shot Add user authentication", "/one-shot Fix navigation bug"],
+    content: `One-Shot Mode (beta) - End-to-end autonomous task execution.
+
+\u26A0\uFE0F  This mode is experimental. Use for straightforward tasks first.
+
+## Workflow
+
+**Phase 1: Requirements Gathering**
+- Interactive selection of task type (Feature, Bug Fix, Refactoring, etc.)
+- Scope clarification
+- Dependencies identification
+- Success criteria definition
+- User selects progress level (Minimal/Moderate/Detailed/Quiet)
+
+**Phase 2: Planning**
+- Delegate to @planner agent
+- Create detailed implementation plan
+- Recommend relevant skills and tools
+- Create Beads task for tracking
+
+**Phase 3: Complexity Check & Auto-Split**
+- Analyze task complexity
+- Split into multiple beads if needed:
+  * Time > 30 minutes
+  * >10 files affected
+  * >500 lines to change
+  * Touches >2 sub-systems
+
+**Phase 4: Execution**
+- Build dependency graph
+- Execute tasks in parallel (max 3 concurrent)
+- Dynamic agent selection (@build \u2192 @review \u2192 @scout \u2192 ...)
+- Integrate skills (TDD, debugging, etc.)
+- Smart terminal access (auto-allow/ask/forbid)
+
+**Phase 5: Enhanced Testing & Validation**
+- Auto-generate test scripts for new functionality
+- Run quality gates: typecheck, test, lint, build
+- Execute sample commands (with user approval)
+- Validate logs semantically with historical comparison
+- Retry loop (max 3 attempts) with:
+  * Auto-fix type errors, lint errors
+  * Alternative approaches from @review
+  * User intervention on final failure
+
+**Phase 6: Multi-Level Verification**
+- All quality gates passed \u2713
+- Manual verification confirmation
+- Deployment approval (if needed)
+- Rollback confirmation (if verification fails)
+
+**Phase 7: Completion**
+- Generate proof of completion
+  * Files changed
+  * Test results
+  * Build output
+  * Deployment status
+- Update Beads task \u2192 completed
+- Store proof in bead notes
+- Collect beta feedback
+
+## Quality Gates (Must ALL Pass)
+- \`npm run typecheck\` - No type errors
+- \`npm run test\` - All tests pass
+- \`npm run lint\` - No linting errors
+- \`npm run build\` - Build succeeds
+
+## Success Criteria
+- All tests passing
+- No regressions
+- Manual verification
+- Deployment complete (if applicable)
+- Beads task completed with proof
+
+## Error Handling
+- **Level 1**: Auto-fix (type errors, lint --fix)
+- **Level 2**: Alternative approach (@review delegation)
+- **Level 3**: User intervention + follow-up bead creation
+
+## Tips
+\u2713 Use for straightforward tasks first
+\u2713 Consider /plan + /implement for complex features
+\u2713 Review changes before final approval`
   },
   // Quick Actions
   {