---

name: investigation-cycle description: HAIOS Investigation Cycle for structured research and discovery. Use when starting or resuming an investigation. Guides HYPOTHESIZE->EXPLORE->CONCLUDE workflow with phase-specific tooling. recipes:

• inv generated: 2025-12-22 last_updated: '2026-01-16T22:00:54'

---

Investigation Cycle

This skill defines the HYPOTHESIZE-EXPLORE-CONCLUDE cycle for structured investigation of research questions. It composes existing primitives (Skills, Commands, Memory) into a coherent research workflow.

When to Use

SHOULD invoke this skill when:

• Starting a new investigation
• Resuming work on an in-progress investigation
• Unsure of next step in research workflow

Invocation: Skill(skill="investigation-cycle")

---

The Cycle

CopyHYPOTHESIZE --> EXPLORE --> CONCLUDE --> CHAIN
      ^            ^           |           |
      |            +-----------+     [route next]
      +-- (if no investigation doc)        |
                                    /-------------\
                                   INV-*    has plan?   else
                                     |         |          |
                                investigation  implement  work-creation
                                   -cycle      -cycle     -cycle

Parallel to Implementation Cycle:

---

1. HYPOTHESIZE Phase

On Entry:

Copyjust set-cycle investigation-cycle HYPOTHESIZE {work_id}

Goal: Verify investigation exists and is ready for exploration.

Actions:

1. Read work file: docs/work/active/{backlog_id}/WORK.md
2. MUST read ALL @ referenced files at document top before proceeding
3. MUST Glob referenced directories to find all files (e.g., @docs/work/active/INV-052/ → Glob("docs/work/active/INV-052/**/*.md"))
4. Verify work file has filled-in Context section (not template placeholders)
5. Check status: active - if draft, fill in context/hypotheses first
6. Query memory for prior related work

Exit Criteria:

• Work file exists with complete context
• MUST: All @ referenced files read (Session 171 learning)
• Hypotheses defined (at least H1)
• Scope defined (in/out)
• Memory queried for prior related findings

Tools: Read, Glob, memory_search_with_experience

---

2. EXPLORE Phase

On Entry:

Copyjust set-cycle investigation-cycle EXPLORE {work_id}

Goal: Execute investigation steps and document findings.

Guardrails (MUST follow):

1. MUST invoke investigation-agent for evidence gathering - specialized subagent handles exploration
2. One hypothesis at a time - Focus exploration
3. Document findings as discovered - Don't wait until end
4. Query memory before assuming - Prior work may answer questions
5. MUST update Exploration Plan checklist after investigation-agent returns - mark all executed steps as complete

Actions:

1. Execute investigation steps from document (via investigation-agent)
2. After subagent returns: Update Exploration Plan checklist to mark completed steps [x]
3. For each step, update Findings section
4. Use memory-agent for context before deep dives
5. If new hypotheses emerge, add them to document

Exit Criteria:

• All investigation steps executed
• Exploration Plan checklist updated (all steps marked [x])
• Findings section populated with evidence
• Hypotheses marked as confirmed/refuted/inconclusive

Tools: Read, Grep, Glob, Bash, memory_search_with_experience, WebSearch (if needed)

---

3. CONCLUDE Phase

On Entry:

Copyjust set-cycle investigation-cycle CONCLUDE {work_id}

Goal: Synthesize findings and spawn work items.

Actions:

1. Review findings against original objective
2. Identify spawned work items (ADRs, backlog items, new investigations)
3. Create spawned items using /new-* commands with spawned_by: {this_investigation_id}
4. Store findings summary to memory via ingester_ingest
5. Update investigation status: status: complete

Exit Criteria:

• Findings synthesized (answer to objective documented)
• Spawned work items created with spawned_by field linking to investigation
• Learnings stored to memory (memory_refs populated)
• Investigation marked complete

Tools: ingester_ingest, Edit, /new-adr, /new-plan

---

4. CHAIN Phase (Post-CONCLUDE)

On Entry:

Copyjust set-cycle investigation-cycle CHAIN {work_id}

Goal: Close investigation and route to next work item.

Actions:

1. Close investigation: /close {backlog_id}
2. Query next work: just ready
3. If items returned, read first work file to check documents.plans
4. Apply routing decision table (see routing-gate skill):
   • If next_work_id is None → await_operator
   • If ID starts with INV- → invoke_investigation
   • If has_plan is True → invoke_implementation
   • Else → invoke_work_creation
5. Execute the action:
   • invoke_investigation -> Skill(skill="investigation-cycle")
   • invoke_implementation -> Skill(skill="implementation-cycle")
   • invoke_work_creation -> Skill(skill="work-creation-cycle")
   • await_operator -> Report "No unblocked work. Awaiting operator direction."

MUST: Do not pause for acknowledgment - execute routing immediately.

Exit Criteria:

• Investigation closed via /close
• Next work item identified (or none available)
• Appropriate cycle skill invoked (or awaiting operator)

On Complete:

Copyjust clear-cycle

Tools: /close, Bash(just ready), Read, Skill(routing-gate)

---

Composition Map

---

Quick Reference

EXPLORE phase guardrails:

• One hypothesis at a time
• Document findings as discovered
• Query memory before assuming

---

Key Design Decisions

---

Related

• implementation-cycle skill: Parallel workflow for implementation tasks
• Investigation template: .claude/templates/investigation.md
• ADR-033: Work Item Lifecycle (DoD criteria)
• /new-investigation command: Creates investigation documents
• /close command: Closes completed investigations