Update ralph-loop to guardrail hooks task

Switch from completed morning-report to guardrail hooks implementation.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

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

.claude/ralph-loop.local.md

@@ -1,25 +1,23 @@
 ---
 active: true
-iteration: 36
+iteration: 1
-max_iterations: 0
+max_iterations: 20
-completion_promise: "The morning-report skill is fully implemented, tested, and registered"
+completion_promise: "Guardrail hooks are fully implemented, tested, and registered"
-started_at: "2026-01-03T08:16:44Z"
+started_at: "2026-01-07T18:52:10Z"
 ---
 
-Build the morning-report skill following the design at ~/.claude/docs/plans/2025-01-02-morning-report-design.md
+Implement guardrail hooks following the design at ~/.claude/docs/plans/2025-01-06-guardrail-hooks-design.md
 
 Implementation order:
-1. Create skill skeleton: ~/.claude/skills/morning-report/ with SKILL.md and config.json
+1. Create state/guardrails.json with starter rules config
-2. Build collectors: weather.py, stocks.py, infra.py (easy wins first)
+2. Create hooks/scripts/guardrail.py main logic
-3. Build gtasks.py collector (Google Tasks API - add OAuth scope)
+3. Create hooks/scripts/guardrail-confirm.py confirm helper
-4. Build news.py collector (RSS feeds)
+4. Modify hooks/hooks.json to add PreToolUse registration
-5. Build generate.py orchestrator and render.py templating
+5. Modify hooks/scripts/session-end.sh to clear session allowlist
-6. Create systemd timer and /morning command
+6. Create logs/ directory
-7. Test end-to-end and verify output
+7. Test: block scenario (catastrophic command pattern)
-
+8. Test: confirm scenario (operation outside safe paths)
-Use appropriate LLM tiers:
+9. Test: allow scenario (operation in safe path)
-- Haiku: weather, stocks, infra formatting
+10. Test: git-aware detection
-- Sonnet: email triage, news summarization
-- None: calendar, tasks (structured data)
 
 Register in component-registry.json when complete.