/DELIVERABLE-SUMMARY

Research Framework Commands - Deliverable Summary

**Date**: 2026-01-25 **Status**: Complete **Owner**: Requirements Analyst **Team**: Research Framework

Summary

Created 10 command definitions and comprehensive documentation for the AIWG Research Framework. Commands cover the complete research lifecycle from discovery to archival, with natural language skill definitions for each.

Deliverables

Command Definitions (10 files)

| File | Lines | Use Case | Description | |------|-------|----------|-------------| | `research-discover.md` | 459 | UC-RF-001 | Semantic search for papers via Semantic Scholar API | | `research-acquire.md` | 347 | UC-RF-002 | Download and validate PDFs from open access sources | | `research-document.md` | 56 | UC-RF-003 | Generate structured notes and extract claims | | `research-cite.md` | 54 | UC-RF-004 | Format citations and back claims with evidence | | `research-archive.md` | 54 | UC-RF-007 | Package and preserve research collections (OAIS) | | `research-quality.md` | 53 | UC-RF-006 | Assess FAIR compliance and quality | | `research-gap-analysis.md` | 52 | UC-RF-009 | Identify research gaps and future directions | | `research-export.md` | 54 | UC-RF-010 | Export to BibTeX, Obsidian, Zotero, OAIS formats | | `research-workflow.md` | 87 | UC-RF-008 | Execute complete end-to-end research pipeline | | `research-status.md` | 93 | Cross-cutting | Monitor framework state and coverage |

**Total**: 1,309 lines of command definitions

Documentation (2 files)

| File | Lines | Description | |------|-------|-------------| | `README.md` | 562 | Comprehensive command reference with workflows, examples, agent mapping | | `DELIVERABLE-SUMMARY.md` | This file | Deliverable documentation |

**Total**: 562+ lines of documentation

Grand Total

**11 files, 1,871+ lines of documentation**

Command Structure

Each command definition includes:

1. **Frontmatter** (YAML metadata)

• Description
• Category
• Argument hints
• Allowed tools
• Model specification

2. **Task Description**

• Purpose and goals
• When to invoke

3. **Parameters**

• Required and optional arguments
• Default values

4. **Inputs/Outputs**

• File paths
• Data formats

5. **Workflow Steps**

• Detailed execution flow
• Code examples (Bash, TypeScript)

6. **Examples**

• Multiple use cases
• Expected outputs

7. **Related Agents**

• Agent mapping
• References to agent specs

8. **Skill Definition**

• Natural language patterns
• Example invocations
• Parameter extraction

9. **Success Criteria**

• Measurable outcomes
• Quality thresholds

10. **Error Handling**

• Common failure scenarios
• Remediation steps

11. **Best Practices**

• Usage recommendations

12. **References**

• Links to use cases, agents, external docs

13. **Security Notes**

• Token security compliance
• API key management

Natural Language Skills

All 10 commands support natural language invocation:

| Command | Natural Language Example | |---------|-------------------------| | /research-discover | "Find papers on OAuth2 security" | | /research-acquire | "Download the papers" | | /research-document | "Summarize the papers" | | /research-cite | "Cite this claim in APA style" | | /research-archive | "Archive the research" | | /research-quality | "Assess research quality" | | /research-gap-analysis | "What's missing in the literature?" | | /research-export | "Export to BibTeX" | | /research-workflow | "Complete research workflow for LLM evaluation" | | /research-status | "Show research status" |

Workflow Coverage

4 Documented Workflows

1. **Quick Research Discovery** (10 minutes, 10 papers)

• discover → select → acquire → document → export

2. **Systematic Literature Review** (30 minutes, 50 papers)

• preregister → discover → acquire → document → quality → gap-analysis → archive → export

3. **Citation-Driven Exploration** (15 minutes, 30 papers)

• discover --citation-network → acquire → build-network → export

4. **End-to-End Automation** (5 minutes, 20 papers)

• workflow (single command, full pipeline)

Agent Mapping

| Agent | Commands | Lines | |-------|----------|-------| | Discovery Agent | /research-discover, /research-gap-analysis | 511 | | Acquisition Agent | /research-acquire | 347 | | Documentation Agent | /research-document | 56 | | Citation Agent | /research-cite | 54 | | Archival Agent | /research-archive, /research-export | 108 | | Quality Agent | /research-quality | 53 | | Workflow Agent | /research-workflow | 87 |

Performance Targets

| Command | Target | NFR Reference | |---------|--------|---------------| | /research-discover | <10 seconds (100 papers) | NFR-RF-D-01 | | /research-discover (gap analysis) | <30 seconds (100 papers) | NFR-RF-D-02 | | /research-acquire | <5 seconds per paper | NFR-RF-A-01 | | /research-document | <30 seconds per paper | NFR-RF-DO-01 | | /research-workflow | <5 minutes (20 papers) | UC-RF-008 |

Integration Points

External APIs

1. **Semantic Scholar API** (Primary discovery)

• Rate limit: 100 req/5 min
• Coverage: 200M+ papers

2. **Unpaywall API** (Open access discovery)

• Polite rate limiting
• Coverage: 30M+ papers

3. **arXiv API** (Preprint access)

• Rate limit: 1 req/3 sec
• Coverage: 2M+ preprints

Export Formats

1. **BibTeX** (Bibliography management) 2. **Obsidian** (Markdown vault with backlinks) 3. **Zotero** (Library import) 4. **OAIS SIP** (Preservation package)

Security Compliance

All commands follow token security rules:

• [x] API keys loaded from `.aiwg/research/config.yaml`, never hardcoded
• [x] Query sanitization to prevent injection attacks
• [x] Rate limiting strictly enforced
• [x] File integrity validation with checksums
• [x] Download from trusted sources only
• [x] No sensitive data logged

**Reference**: @$AIWG_ROOT/agentic/code/frameworks/sdlc-complete/rules/token-security.md

Traceability

All