# Workflow Tt Research > Execute a research job with auto-strategy selection using the timetravel adapter library. Invoke with /workflow tt-research. - Skill: `lev-os/workflow-tt-research` (Agent Skill) - Install (CLI): `npx skillmds@latest add lev-os/workflow-tt-research` - Raw SKILL.md: https://api.skillmd.com/api/skills/lev-os/workflow-tt-research/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Research & Search - Author: lev-os (https://skillmd.com/u/lev-os) - Updated: 2026-09-10 - Page: https://skillmd.com/skills/lev-os/workflow-tt-research --- # Workflow: Timetravel Research Execute a full research workflow using the timetravel adapter library at `~/digital/leviathan/plugins/timetravel/`. Every step spawns a **fresh subagent** (poly sparkle pattern). State flows only through handoff artifacts at `.lev/pm/handoffs/tt-research-*.md`. ## Trigger - User wants to research a topic using the timetravel system - `/workflow tt-research "" [--strategy ] [--depth quick|standard|deep]` ## Inputs | Input | Required | Description | |-------|----------|-------------| | `query` | yes | Research query string | | `strategy` | no | Override strategy (quick, full, deep, max, academic, social). Default: auto-selected | | `depth` | no | Shorthand for strategy: quick=quick, standard=full, deep=deep | | `output_dir` | no | Override output directory (default: `~/.config/LEV/research/sessions///`) | ## Steps ### Step 1: Assess Query **Agent Type**: Explore **Skills**: `skill://lev-research` **Research**: - Analyze query complexity: single-fact, comparative, investigative, comprehensive - Read available strategies: `~/digital/leviathan/plugins/timetravel/src/strategy.ts` - Check user-provided strategy override **Action**: - Classify query type and recommend strategy - If no strategy specified: quick for single-fact, full for comparative, deep for investigative, max for comprehensive - Identify relevant adapters and potential gaps - Create session directory: `~/.config/LEV/research/sessions///` **Handoff**: `.lev/pm/handoffs/tt-research-1-assess.md` ```yaml query: query_type: single-fact | comparative | investigative | comprehensive recommended_strategy: selected_strategy: session_dir: adapters_needed: [] ``` --- ### Step 2: Health Check **Agent Type**: general-purpose **Skills**: None **Research**: - Read handoff from Step 1 - Run `cd ~/digital/leviathan/plugins/timetravel && npx timetravel status` to check adapter availability **Action**: - Verify all adapters needed by selected strategy are available - If missing adapters: downgrade strategy or warn user - Record adapter health snapshot **Handoff**: `.lev/pm/handoffs/tt-research-2-health.md` ```yaml strategy_viable: true | false available_adapters: [] missing_adapters: [] adjusted_strategy: ``` --- ### Step 3: Execute Search **Agent Type**: general-purpose **Skills**: None **Research**: - Read handoffs from Steps 1-2 - Read strategy definition for orchestration mode **Action**: - Run `cd ~/digital/leviathan/plugins/timetravel && npx timetravel search "" --strategy ` - Capture JSON output - Save raw results to `/artifacts/search_raw.json` - If strategy has synthesizer, verify synthesis was included **Handoff**: `.lev/pm/handoffs/tt-research-3-execute.md` ```yaml strategy_used: adapters_responded: [] source_count: has_synthesis: true | false raw_results_path: duration_ms: ``` --- ### Step 4: Synthesize **Agent Type**: general-purpose **Skills**: `skill://lev-research` **Research**: - Read handoff from Step 3 - Read raw results from `/artifacts/search_raw.json` - If synthesis already present from oracle, use as base **Action**: - Deduplicate sources across adapters (by URL) - Rank sources by relevance (title/snippet match to query) - Generate synthesis: executive summary (2-3 sentences), key findings (3-5), evidence table - Apply multi-perspective overlay if depth is deep or max - Save to `/synthesis.md` **Handoff**: `.lev/pm/handoffs/tt-research-4-synthesize.md` ```yaml unique_sources: key_findings: perspectives_applied: [] synthesis_path: confidence: <0.0-1.0> ``` --- ### Step 5: Generate Report **Agent Type**: general-purpose **Skills**: None **Research**: - Read handoffs from Steps 1-4 - Read synthesis from Step 4 **Action**: - Compile final research report using lev-research synthesis template: - Executive summary - Key findings with source citations - Evidence table (claim, source, confidence) - Open questions - Numbered source list with URLs - Save to `/report.md` - Save `/session.json` with metadata **Handoff**: `.lev/pm/handoffs/tt-research-5-report.md` ```yaml report_path: session_json_path: total_sources: confidence: <0.0-1.0> ``` --- ### Step 6: Persist Job **Agent Type**: general-purpose **Skills**: None **Research**: - Read handoffs from Steps 1-5 - Read job system: `~/digital/leviathan/plugins/timetravel/src/job.ts` **Action**: - Create a timetravel job record for re-runs: `npx timetravel job create -q "" -s ` - Record job ID in session metadata - Optionally schedule if user requested recurring: `npx timetravel schedule add -c ` - Print final summary with report path and job ID **Handoff**: `.lev/pm/handoffs/tt-research-6-persist.md` ```yaml job_id: scheduled: true | false schedule_interval: report_path: status: complete ``` ## Outputs - Research report: `/report.md` - Synthesis: `/synthesis.md` - Raw results: `/artifacts/search_raw.json` - Session metadata: `/session.json` - Job record: `~/.config/LEV/timetravel/jobs//` - 6 handoff artifacts in `.lev/pm/handoffs/tt-research-*.md` ## Validation - [ ] Search returned at least 1 source - [ ] Synthesis includes executive summary and key findings - [ ] Report follows lev-research template structure - [ ] Session directory contains all expected files - [ ] Job record created and retrievable via `timetravel job list` - [ ] All 6 handoff files created ## Usage ```bash /workflow tt-research "state of AI agent frameworks 2026" /workflow tt-research "rust vs go for CLI tools" --strategy academic /workflow tt-research "trending AI on twitter" --strategy social /workflow tt-research "comprehensive LLM evaluation" --depth deep ``` ## Technique Map - **Role definition** - Clarifies operating scope and prevents ambiguous execution. - **Context enrichment** - Captures required inputs before actions. - **Output structuring** - Standardizes deliverables for consistent reuse. - **Step-by-step workflow** - Reduces errors by making execution order explicit. - **Edge-case handling** - Documents safe fallbacks when assumptions fail. ## Technique Notes These techniques improve reliability by making intent, inputs, outputs, and fallback paths explicit. Keep this section concise and additive so existing domain guidance remains primary. ## Prompt Architect Overlay ### Role Definition You are the prompt-architect-enhanced specialist for workflow-tt-research, responsible for deterministic execution of this skill's guidance while preserving existing workflow and constraints. ### Input Contract - Required: clear user intent and relevant context for this skill. - Preferred: repository/project constraints, existing artifacts, and success criteria. - If context is missing, ask focused questions before proceeding. ### Output Contract - Provide structured, actionable outputs aligned to this skill's existing format. - Include assumptions and next steps when appropriate. - Preserve compatibility with existing sections and related skills. ### Edge Cases & Fallbacks - If prerequisites are missing, provide a minimal safe path and request missing inputs. - If scope is ambiguous, narrow to the highest-confidence sub-task. - If a requested action conflicts with existing constraints, explain and offer compliant alternatives.