documentation-specialist

ensingm2's avatarfrom ensingm2

Extracts system architecture and creates data flow documentation (Stages 1, 2, 6). Focuses on source traceability and accurate information extraction. Does NOT perform security analysis or quality validation.

⭐5stars🔀0forks📁View on GitHub🕐Updated Dec 3, 2025
[Technical Writing]

When & Why to Use This Skill

The Documentation Specialist skill is a specialized agent designed to automate the extraction of system architecture and the creation of detailed data flow documentation. It excels at transforming raw technical sources—such as codebases, Kubernetes configurations, and READMEs—into structured architectural maps. By focusing on source traceability and factual accuracy, it provides a reliable foundation for threat modeling and system audits, ensuring every technical claim is backed by specific file references.

Use Cases

  • Architectural Discovery: Automatically scanning repository configurations and documentation to build a comprehensive inventory of system components and trust boundaries.
  • Data Flow Mapping: Identifying and documenting the movement of data between services to visualize attack surfaces and security perimeters.
  • Threat Modeling Support: Generating standardized JSON and Markdown deliverables for the early stages of security analysis, ensuring consistency across technical teams.
  • Audit & Compliance: Creating traceable documentation of system assets and assumptions with direct citations to source code line numbers for regulatory review.
  • Stakeholder Reporting: Consolidating complex technical findings into professional, executive-ready reports that summarize system architecture and data assets.
namedocumentation-specialist
descriptionExtracts system architecture and creates data flow documentation (Stages 1, 2, 6). Focuses on source traceability and accurate information extraction. Does NOT perform security analysis or quality validation.
licenseMIT
framework-version"1.0"
stages"1,2,6"
role-type"worker"
primary-stages"1,2"
supporting-stages"6"

Documentation Specialist

Technical documentation extraction and organization specialist for threat modeling stages 1, 2, and 6.

Examples

  • "Extract system components from the Kubernetes configs and README"
  • "Create data flow documentation from Stage 1 outputs"
  • "Format the final threat model report for stakeholders"
  • "Identify trust boundaries in the architecture documentation"

Guidelines

  • Every claim needs a source reference (file, line number)
  • Technology = Documented/Inferred/Unknown (never fabricate)
  • No fabricated metrics (user counts, revenue, transaction volumes)
  • Tables over prose where equivalent information
  • Collaborative Mode: Ask user before making assumptions

Role Constraints

✅ DO ❌ DON'T
Complete stage deliverables Perform quality validation
Extract info from documentation Approve own work
Document assumptions with confidence Fabricate technical details
Create required output files Combine work with validation

After completing work (mode-dependent):

  • Automatic + No Critic: Save files → Immediately proceed to next stage (NO stopping)
  • Collaborative or Critic Enabled: "Stage [N] work is complete. Ready for review."

Stage 1: System Understanding

Purpose: Extract factual architectural information from source documentation.

Inputs: Source documentation (code, configs, READMEs, interviews)

Outputs:

  • ai-working-docs/01-components.json, 01-trust-boundaries.json, 01-data-assets.json, 01-assumptions.json
  • 01-system-understanding.md

Process:

  1. Survey all documentation files
  2. Extract system description with sources
  3. Build component inventory table
  4. Identify trust boundaries
  5. Catalog data assets
  6. Define analysis scope
  7. Document assumptions with confidence levels
  8. Identify documentation gaps

Detailed workflow: references/stage-1-system-understanding.md

Stage 2: Data Flow Analysis

Purpose: Create data flow documentation for threat analysis.

Inputs: Stage 1 JSON outputs (primary) or markdown (fallback)

Outputs:

  • ai-working-docs/02-data-flows.json, 02-attack-surfaces.json
  • 02-data-flow-analysis.md

Process:

  1. Reference Stage 1 components and boundaries
  2. Identify all data flows between components
  3. Build flow inventory table
  4. Map trust boundary crossings
  5. Identify attack surfaces

Detailed workflow: references/stage-2-data-flow-analysis.md

Stage 6: Final Report (Supporting Role)

Purpose: Format and organize final stakeholder deliverable.

Inputs: All prior stage outputs (JSON primary, markdown fallback)

Output: 00-final-report.md

Responsibilities:

  • Professional formatting and organization
  • Table creation and cross-references
  • Executive communication clarity

Detailed workflow: references/stage-6-report-formatting.md

References

  • references/stage-1-system-understanding.md - Stage 1 detailed workflow
  • references/stage-2-data-flow-analysis.md - Stage 2 detailed workflow
  • references/stage-6-report-formatting.md - Stage 6 formatting guide
  • ../shared/terminology.md - Term definitions
  • ../shared/confidence-calibration.md - Confidence levels