docs: remove NLU from v1.0 scope

- Remove Claude API integration and intent parsing (04-02-PLAN)
- REQ-08 (conversational queries) moved to out of scope
- Phase 4 renamed from "Logs & Intelligence" to "Logs" (complete)
- v1.0 now focuses on keyword-based container control

Simple substring matching works well for container management.
NLU adds complexity without proportional value for v1.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

.planning/phases/04-logs/04-01-SUMMARY.md

@@ -0,0 +1,130 @@
+---
+phase: 04-logs-intelligence
+plan: 01
+subsystem: logs
+tags: [docker-logs, telegram-bot, n8n, log-streaming]
+
+# Dependency graph
+requires:
+  - phase: 02-docker-integration
+    provides: Docker API integration and container matching patterns
+  - phase: 03-container-actions
+    provides: Action command routing and fuzzy container matching
+provides:
+  - Container log retrieval via Telegram commands
+  - Configurable line count (default 50, max 1000)
+  - Binary stream decoder for Docker logs API
+  - Formatted log output with monospace display
+affects: [04-02-claude-analysis]
+
+# Tech tracking
+tech-stack:
+  added: []
+  patterns:
+    - "Docker logs API multiplexed stream decoding"
+    - "Telegram message truncation for long content"
+
+key-files:
+  created: []
+  modified:
+    - n8n-workflow.json
+
+key-decisions:
+  - "Use Docker logs API with tail parameter for line-limited retrieval"
+  - "Strip 8-byte binary headers from Docker multiplexed stream"
+  - "Truncate at 3800 chars to stay within Telegram 4096 limit"
+  - "Default to 50 lines when not specified, cap at 1000 max"
+
+patterns-established:
+  - "Binary stream decoding: check charCodeAt(0) <= 2 for header presence"
+  - "Log formatting: <pre> tags for monospace, HTML parse mode"
+
+# Metrics
+duration: 3min
+completed: 2026-01-30
+---
+
+# Phase 04 Plan 01: Container Log Retrieval Summary
+
+**Container log retrieval via Telegram with Docker API binary stream decoding and configurable line limits**
+
+## Performance
+
+- **Duration:** 3 min
+- **Started:** 2026-01-31T02:40:38Z
+- **Completed:** 2026-01-31T02:43:37Z
+- **Tasks:** 3
+- **Files modified:** 1
+
+## Accomplishments
+- Users can request logs for any container by name via Telegram
+- Configurable line count (default 50, max 1000 lines)
+- Docker logs API binary stream format properly decoded
+- Logs displayed in readable monospace format with truncation for Telegram limits
+
+## Task Commits
+
+Implementation was atomic - all tasks completed in single commit:
+
+1. **Tasks 1-3: Complete logs implementation** - `93c40fe` (feat)
+   - Task 1: Logs command routing
+   - Task 2: Docker logs API integration
+   - Task 3: Binary stream format handling
+
+**Note:** All three tasks were interdependent and completed together in a single atomic commit since they form a cohesive feature implementation.
+
+## Files Created/Modified
+- `n8n-workflow.json` - Added 11 new nodes for complete logs workflow:
+  - Parse Logs Command: Extract container name and line count from user message
+  - Docker List for Logs: Fetch containers for matching
+  - Match Logs Container: Fuzzy match container name
+  - Check Logs Match Count: Route based on match results
+  - Build Logs Command: Construct Docker API curl with tail parameter
+  - Execute Logs: Call Docker logs API
+  - Format Logs: Decode binary stream, truncate, add monospace formatting
+  - Send Logs Response: Reply to user via Telegram
+  - Error handlers: No match, multiple matches, Docker errors
+
+## Decisions Made
+
+**1. Docker logs API with tail parameter**
+- **Rationale:** More efficient than fetching all logs and filtering in code
+- **Implementation:** `http://localhost/v1.47/containers/{id}/logs?stdout=1&stderr=1&tail={lines}&timestamps=1`
+
+**2. Binary stream header detection via charCodeAt(0)**
+- **Rationale:** Docker multiplexed stream uses byte 0 values 1 (stdout) or 2 (stderr)
+- **Implementation:** Check if `charCodeAt(0) <= 2` and length > 8, then strip first 8 bytes
+- **Robustness:** Works even if some lines lack headers (mixed format tolerance)
+
+**3. Default 50 lines, cap at 1000**
+- **Rationale:** Balance between useful context and Telegram message limits
+- **Safety:** 1000 line cap prevents excessive API calls and message truncation
+
+**4. Truncate at 3800 chars**
+- **Rationale:** Telegram limit is 4096, leave room for header and formatting tags
+- **UX:** Add "... (truncated)" indicator when limit hit
+
+## Deviations from Plan
+
+None - plan executed exactly as written.
+
+## Issues Encountered
+
+None - Docker logs API worked as expected, binary stream format handled correctly.
+
+## User Setup Required
+
+None - no external service configuration required.
+
+## Next Phase Readiness
+
+Ready for Phase 04 Plan 02 (Claude log analysis):
+- Log retrieval working and tested
+- Format suitable for Claude API input
+- Error handling in place for edge cases
+
+Logs are now available programmatically for Claude to analyze in next plan.
+
+---
+*Phase: 04-logs-intelligence*
+*Completed: 2026-01-30*