UNPKG

aiwg

Version:

Deployment tool and support utility for AI context. Copies agents, skills, commands, rules, and behaviors into the paths each AI platform reads (Claude Code, Codex, Copilot, Cursor, Warp, OpenClaw, and 6 more) so one source of truth works across 10 platfo

386 lines (298 loc) 13.2 kB
# Gap Analysis Report ## Metadata ```yaml --- report_type: gap-analysis report_date: YYYY-MM-DD corpus_size: 0 # Number of sources analyzed coverage_score: 0 # 0-100 overall coverage identified_gaps: 0 # Total gaps found high_priority_gaps: 0 medium_priority_gaps: 0 low_priority_gaps: 0 analyst: gap-analysis-agent version: 1.0.0 status: draft # draft | final | archived --- ``` ## Executive Summary **Corpus Coverage:** X% (Target: 100%) **Gaps Identified:** X total (Y high priority, Z medium, W low) **Recommendation:** [Fill high-priority gaps | Coverage sufficient | Expand corpus] **Key Findings:** - [Most critical gap] - [Second critical gap] - [Notable insight about corpus composition] --- ## Corpus Overview ### Size and Composition | Metric | Count | Percentage | |--------|-------|------------| | **Total Sources** | 0 | 100% | | Academic Papers | 0 | 0% | | Preprints | 0 | 0% | | Blog Posts | 0 | 0% | | Documentation | 0 | 0% | | Books/Chapters | 0 | 0% | ### Date Range Distribution | Year Range | Count | Percentage | |------------|-------|------------| | 2024-2026 | 0 | 0% | | 2022-2023 | 0 | 0% | | 2020-2021 | 0 | 0% | | 2019 or earlier | 0 | 0% | ### Quality Distribution | GRADE Rating | Count | Percentage | |--------------|-------|------------| | High | 0 | 0% | | Moderate | 0 | 0% | | Low | 0 | 0% | | Very Low | 0 | 0% | | Not Assessed | 0 | 0% | --- ## Coverage Analysis ### Topic Coverage #### Expected Topics (Based on Taxonomy) | Topic | Expected Coverage | Actual Coverage | Gap Size | Priority | |-------|------------------|-----------------|----------|----------| | Multi-agent frameworks | 30% (7-8 sources) | 48% (12 sources) | Exceeded | - | | Prompt engineering | 25% (6-7 sources) | 32% (8 sources) | Met | - | | Agent memory systems | 20% (5 sources) | 12% (3 sources) | ⚠️ Under-represented | Medium | | Agent tool use | 15% (3-4 sources) | 8% (2 sources) | ⚠️ Under-represented | Medium | | **Agent safety** | **10% (2-3 sources)** | **0% (0 sources)** | **❌ Missing** | **High** | | **Evaluation methods** | **10% (2-3 sources)** | **0% (0 sources)** | **❌ Missing** | **High** | **Overall Topic Coverage:** 66% (4/6 expected topics covered) #### Topic Coverage Gaps Identified ##### GAP-T-001: Agent Safety (Missing Topic) - **Expected Coverage:** 2-3 sources (10%) - **Actual Coverage:** 0 sources (0%) - **Gap Size:** 100% missing - **Priority:** High - **Impact:** Critical for production AIWG deployment - **Recommendation:** Acquire 3-5 sources on AI safety, agent alignment, constitutional AI - **Search Query:** `"AI safety" OR "agent alignment" OR "constitutional AI"` - **Expected Sources:** REF-025 (Constitutional AI), safety frameworks, guardrails research ##### GAP-T-002: Evaluation Methods (Missing Topic) - **Expected Coverage:** 2-3 sources (10%) - **Actual Coverage:** 0 sources (0%) - **Gap Size:** 100% missing - **Priority:** High - **Impact:** Cannot validate AIWG agent performance without evaluation framework - **Recommendation:** Acquire 2-3 sources on agent benchmarking, task performance metrics - **Search Query:** `"agent evaluation" OR "benchmarking" OR "task performance metrics"` ##### GAP-T-003: Agent Tool Use (Under-Represented) - **Expected Coverage:** 3-4 sources (15%) - **Actual Coverage:** 2 sources (8%) - **Gap Size:** 47% under-represented - **Priority:** Medium - **Impact:** Limited evidence for tool integration patterns - **Recommendation:** Acquire 1-2 additional sources on function calling, API usage - **Search Query:** `"agent tool use" OR "function calling" OR "API integration"` --- ### Citation Completeness #### Citation Network Analysis | Metric | Count | Status | |--------|-------|--------| | **Total Citations** | 245 | - | | **In Corpus** | 233 | | | **Orphaned Citations** | 12 | ⚠️ | | **Citation Completeness** | 95% | Good | #### Orphaned Citations (Frequently Cited, Not in Corpus) ##### GAP-C-001: Constitutional AI Paper (5 citations, not in corpus) - **Title:** "Constitutional AI: Harmlessness from AI Feedback" - **Authors:** Bai, Y., et al. (Anthropic) - **Year:** 2022 - **Cited By:** REF-012, REF-016, REF-021, REF-023, REF-024 - **Priority:** High - **Rationale:** Cited by 20% of corpus (5/25 papers), foundational for AI safety - **Action:** Acquire as REF-025 ##### GAP-C-002: Chain-of-Thought Prompting (4 citations, not in corpus) - **Title:** "Chain-of-Thought Prompting Elicits Reasoning in Large Language Models" - **Authors:** Wei, J., et al. (Google) - **Year:** 2022 - **Cited By:** REF-016, REF-017, REF-021, REF-023 - **Priority:** High - **Rationale:** Foundational for prompt engineering approaches - **Action:** Acquire as REF-026 ##### GAP-C-003: ReAct Framework (3 citations, not in corpus) - **Title:** "ReAct: Synergizing Reasoning and Acting in Language Models" - **Authors:** Yao, S., et al. - **Year:** 2022 - **Cited By:** REF-018, REF-023, REF-024 - **Priority:** Medium - **Rationale:** Influential for reasoning + action agents - **Action:** Acquire as REF-027 [... list all 12 orphaned citations with priority ...] --- ### Source Diversity #### Publication Type Diversity | Type | Count | Percentage | Target | Gap | |------|-------|------------|--------|-----| | Academic Papers | 18 | 72% | 60-70% | Met | | Blog Posts | 5 | 20% | 10-20% | Met | | Documentation | 2 | 8% | 5-10% | Met | | **Books/Chapters** | **0** | **0%** | **5-10%** | **❌ Missing** | ##### GAP-D-001: No Books or Book Chapters - **Expected:** 1-2 sources (5-10%) - **Actual:** 0 sources (0%) - **Priority:** Low - **Impact:** Missing foundational/comprehensive treatments - **Recommendation:** Acquire 1-2 authoritative books on multi-agent systems or LLM applications - **Search Query:** `"multi-agent systems" book OR "LLM applications" book` #### Date Range Analysis | Range | Count | Percentage | Assessment | |-------|-------|------------|------------| | 2024-2026 | 15 | 60% | Good (recent research) | | 2022-2023 | 8 | 32% | Good (recent history) | | 2020-2021 | 2 | 8% | ⚠️ Limited | | **2019 or earlier** | **0** | **0%** | **❌ Missing foundational work** | ##### GAP-D-002: No Foundational Papers (2019 or Earlier) - **Expected:** 1-2 sources (5%) - **Actual:** 0 sources (0%) - **Priority:** Medium - **Impact:** Lacks historical context, may miss fundamental concepts - **Recommendation:** Acquire 2-3 foundational papers on multi-agent systems (pre-LLM era) - **Search Query:** `"multi-agent systems" AND year:2015-2019` #### Author Diversity | Metric | Count | Assessment | |--------|-------|------------| | **Unique Authors** | 45 | Good diversity | | **Top Author Concentration** | 10 sources (40%) | ⚠️ High concentration | | **Unique Organizations** | 15 | Good diversity | ##### GAP-D-003: Author Concentration (Top 3 Authors in 40% of Sources) - **Top Authors:** [Author A: 5 sources, Author B: 3 sources, Author C: 2 sources] - **Priority:** Low - **Impact:** May indicate limited perspective diversity - **Recommendation:** Seek sources from different research groups - **Action:** Review author affiliation distribution in next gap analysis --- ## Coverage Score Calculation ### Dimension Scores | Dimension | Weight | Score | Weighted Score | |-----------|--------|-------|----------------| | **Topic Coverage** | 50% | 66/100 | 33.0 | | **Citation Completeness** | 30% | 95/100 | 28.5 | | **Source Diversity** | 20% | 75/100 | 15.0 | **Overall Coverage Score:** 76.5 / 100 **Interpretation:** - **70-79:** Good coverage with room for improvement - **Target:** 85+ (comprehensive coverage) --- ## Recommendations ### High Priority (Address Before Construction Phase) #### Recommendation 1: Fill "Agent Safety" Topic Gap - **Action:** Acquire 3-5 sources on AI safety, agent alignment, constitutional AI - **Search Query:** `"AI safety" OR "agent alignment" OR "constitutional AI"` - **Target Sources:** - REF-025: Constitutional AI (Anthropic) - Safety frameworks for LLM applications - Guardrails and monitoring approaches - **Estimated Time:** 2 hours (discovery + acquisition + documentation) - **Impact:** +8% topic coverage - **Owner:** Discovery Agent Acquisition Agent Documentation Agent #### Recommendation 2: Fill "Evaluation Methods" Topic Gap - **Action:** Acquire 2-3 sources on agent evaluation, benchmarking - **Search Query:** `"agent evaluation" OR "benchmarking" OR "task performance metrics"` - **Target Sources:** - Agent benchmarking papers (HELM, BIG-Bench) - Human evaluation frameworks - Automatic evaluation metrics - **Estimated Time:** 1.5 hours - **Impact:** +6% topic coverage - **Owner:** Discovery Agent #### Recommendation 3: Acquire Orphaned Citations (Top 3) - **Action:** Add Constitutional AI, Chain-of-Thought, ReAct papers - **Search Query:** Exact titles (known works) - **Estimated Time:** 1 hour (known sources, direct acquisition) - **Impact:** +3% citation completeness (12 9 orphaned) - **Owner:** Acquisition Agent ### Medium Priority (Address During Construction Phase) #### Recommendation 4: Add Foundational Papers (2019 or Earlier) - **Action:** Acquire 2-3 foundational multi-agent systems papers - **Search Query:** `"multi-agent systems" AND year:2015-2019` - **Rationale:** Establish historical context - **Estimated Time:** 2 hours - **Impact:** +8% date diversity - **Owner:** Discovery Agent #### Recommendation 5: Address Under-Represented Topics - **Topics:** Agent tool use (needs 1-2 more sources), Agent memory (needs 2 more) - **Estimated Time:** 1.5 hours total - **Impact:** +10% topic coverage - **Owner:** Discovery Agent ### Low Priority (Post-v1.0 Enhancement) #### Recommendation 6: Add Book Sources - **Action:** Acquire 1-2 authoritative books on multi-agent systems - **Rationale:** Comprehensive treatments for reference - **Estimated Time:** 3 hours (books require careful review) - **Impact:** +8% publication diversity - **Owner:** Manual review + Acquisition #### Recommendation 7: Complete Citation Metadata - **Action:** Document reference lists for 5 sources missing citation metadata - **Estimated Time:** 30 minutes - **Impact:** Citation network completeness - **Owner:** Documentation Agent --- ## Gap-Filling Action Plan ### Immediate Actions (Next 7 Days) | Action | Priority | Owner | Time | Impact | |--------|----------|-------|------|--------| | Acquire Agent Safety sources | High | Discovery Agent | 2h | +8% topic | | Acquire Evaluation sources | High | Discovery Agent | 1.5h | +6% topic | | Acquire orphaned citations (top 3) | High | Acquisition Agent | 1h | +3% citation | **Total Time Investment:** 4.5 hours **Expected Coverage Improvement:** 76.5% 85%+ (Target achieved) ### Follow-Up Actions (Next 30 Days) | Action | Priority | Owner | Time | Impact | |--------|----------|-------|------|--------| | Acquire foundational papers | Medium | Discovery Agent | 2h | +8% date diversity | | Fill under-represented topics | Medium | Discovery Agent | 1.5h | +10% topic | ### Future Enhancements (Post-v1.0) | Action | Priority | Owner | Time | Impact | |--------|----------|-------|------|--------| | Add book sources | Low | Manual | 3h | +8% pub diversity | | Complete citation metadata | Low | Documentation Agent | 30min | Citation network | --- ## Gap Analysis Changelog ### 2026-01-25 (Initial Analysis) - **Corpus Size:** 25 sources - **Coverage Score:** 76.5% - **Gaps Identified:** 8 total (3 high, 3 medium, 2 low) - **Action:** Begin high-priority gap-filling ### [Future Entry Template] - **Date:** YYYY-MM-DD - **Corpus Size:** [Updated size] - **Coverage Score:** [Updated score] - **Changes:** [What was added/removed] - **Gaps Resolved:** [Which gaps were filled] - **New Gaps:** [Any new gaps identified] --- ## Validation Rules ### Required Metrics - `corpus_size`: Integer >0 - `coverage_score`: Integer 0-100 - `identified_gaps`: Integer ≥0 - `report_date`: ISO 8601 date ### Coverage Thresholds - **Excellent:** 85-100% (comprehensive) - **Good:** 70-84% (sufficient for most purposes) - **Fair:** 50-69% (needs improvement) - **Poor:** <50% (significant gaps) ### Gap Priority Criteria - **High:** Missing expected topics (0% coverage), >10 orphaned citations - **Medium:** Under-represented topics (<50% expected coverage), 5-10 orphaned citations - **Low:** Minor diversity gaps, metadata incompleteness --- ## Agent Responsibilities **Produced by:** Gap Analysis Agent (UC-RF-009) **Updated by:** Gap Analysis Agent (periodic re-analysis) **Used by:** Discovery Agent (gap-filling search), User (strategic planning), Workflow Agent (phase gate decisions) --- ## References - @$AIWG_ROOT/agentic/code/frameworks/research-complete/elaboration/use-cases/UC-RF-009-perform-gap-analysis.md - Gap analysis use case - @.aiwg/research/sources/metadata/ - Source metadata for analysis - @.aiwg/research/knowledge/extractions/ - Citation network data - @.aiwg/research/config/topic-taxonomy.yaml - Expected topic coverage (if defined) --- **Template Version:** 1.0.0 **Last Updated:** 2026-01-25 **Owner:** Gap Analysis Agent