docs/specs/README.md

Implementation Specs

Last updated: 2025-12-03

Note: Implemented specs archived to docs/specs/archive/

---

Active Specs (Not Yet Implemented)

SPEC_13: Evidence Deduplication

File: SPEC_13_EVIDENCE_DEDUPLICATION.md Status: Draft - Ready for Implementation Priority: P1 GitHub Issue: #94

Problem: Same paper appears 2-3 times because Europe PMC and OpenAlex both index PubMed.

Solution: Extract PMID from OpenAlex metadata, deduplicate by paper ID in SearchHandler.

---

SPEC_14: ClinicalTrials Outcomes

File: SPEC_14_CLINICALTRIALS_OUTCOMES.md Status: Draft - Ready for Implementation Priority: P1 GitHub Issue: #95

Problem: ClinicalTrials tool misses critical efficacy data (primary outcomes, results availability).

Solution: Add OutcomesModule and HasResults to API fields, extract outcome measures.

---

Implemented Specs (Archived)

All implemented specs are in docs/specs/archive/. Summary:

Spec	Description	PR/Commit
SPEC_01	Demo Termination	Archived
SPEC_02	E2E Testing	Archived
SPEC_03	OpenAlex Integration	Archived
SPEC_04	Magentic UX	Archived
SPEC_05	Orchestrator Cleanup	Archived
SPEC_06	Simple Mode Synthesis	Archived (deleted simple.py)
SPEC_07	LangGraph Memory Arch	Archived
SPEC_08	Integrate Memory Layer	Archived
SPEC_09	LlamaIndex Integration	Archived
SPEC_10	Domain Agnostic Refactor	Archived
SPEC_11	Sexual Health Focus	Archived
SPEC_12	Narrative Synthesis	Archived
SPEC_15	Advanced Mode Performance	PR #65 - max_rounds=5, early termination
SPEC_16	Unified Chat Client Architecture	PR #115 - HuggingFace + OpenAI factory
SPEC_17	Accumulator Pattern	PR #117 - Fixes repr bug

---

How to Write a Spec

1. Create docs/specs/SPEC_{N}_{NAME}.md
2. Include: Problem, Solution, Test Plan, Acceptance Criteria
3. Link to GitHub issue
4. Update this README
5. When implemented, move to archive/