# Visual Source Artifact > Shape product, business, data, research, UX, and mockup artifacts as agent-readable sources with generated visual human review surfaces. - **Type:** Skill - **Install:** `agentstack add skill-45ck-skill-harness-visual-source-artifact` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [45ck](https://agentstack.voostack.com/s/45ck) - **Installs:** 0 - **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [45ck](https://github.com/45ck) - **Source:** https://github.com/45ck/skill-harness/tree/main/packs/developer-artifact-skills/skills/visual-source-artifact - **Website:** https://github.com/45ck/skill-harness/releases/tag/v0.1.0 ## Install ```sh agentstack add skill-45ck-skill-harness-visual-source-artifact ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Visual Source Artifact Use this skill when a human needs to inspect product, business, data, research, UX, or mockup work visually while agents need durable source they can diff, update, and verify. ## Core Rule Use a source-first pair: - Canonical source: Markdown, TOON, JSON, YAML, specgraph-compatible docs, model text, or structured data. - Generated review surface: static HTML, dashboard, state board, prototype, journey map, schema map, evidence board, or comparison page under `generated/review/`. - For human-facing discovery and planning, the generated review surface should be infographic-style HTML under `generated/review//`. Generated visual files are never the only durable source. Edit source first, then regenerate or discard the review surface. ## Default Fidelity - High-fidelity is the default for UI, product, customer-facing workflow, and mockup review. - Low-fidelity sketches are scratch artifacts only unless the project explicitly records them as research evidence. - Use realistic states, dense data, error paths, accessibility cues, assumptions, and evidence strength in the review surface. - Use scan-friendly metrics, charts, timelines, evidence/freshness panels, and source links for non-UI human review artifacts. - Use Mermaid for authored diagrams, Vega-Lite as the default chart grammar, Observable Plot for compact exploratory charts, D3 for bespoke static layouts, Graphviz for node-edge maps, ECharts only as generation-time/static output, RAWGraphs for exported design-led SVGs, and Chart.js only through server-rendered/static output or an equivalent static chart. - Put chart and graph intent in `artifact-infographic` JSON fences or manifest `infographics` arrays when the project has the generic renderer. ## Artifact Families - Product: PRD, opportunity brief, feature map, roadmap, acceptance matrix. - Business: business model, pricing assumptions, stakeholder map, operating risk, go-to-market assumptions. - Data: schema, data dictionary, metric definition, lineage, quality rule, sample records. - Research: claim-evidence matrix, literature theme map, interview synthesis, assumption register. - UX: design brief, component state spec, interaction flow, high-fidelity prototype source, journey map. ## Agent Team - `requirements-analyst` owns product intent, requirements, and acceptance criteria. - `delivery-manager` owns business viability, stakeholder impact, rollout, and risk. - `backend-engineer` owns data shape, schemas, integrity, and metrics implementation constraints. - `research-writer` owns citations, claim strength, research synthesis, and gaps. - `ux-researcher` owns task evidence, prototype critique, high-fidelity UX review, and accessibility concerns. - `system-modeler` owns UML/UWE/C4/workflow impact when structure or behavior changes. - `quality-reviewer` owns freshness, manifest completeness, evidence sufficiency, and readiness risks. ## Process 1. Name the decision the visual artifact supports. 2. Select the artifact family and owner agent. 3. Define the canonical source path under `docs/artifacts/source//` unless a domain docs path is better. 4. Define the generated review path under `generated/review//`. 5. List evidence links and label real evidence separately from synthetic user or agent-simulation evidence. 6. Add or update the manifest entry with source, review surface, owner, evidence, status, freshness, and `reviewRequired: true` when a generated HTML surface is expected. 7. Add `artifact-infographic` specs for charts, graphs, timelines, or source-backed infographic panels when useful. 8. Run `node scripts/generate-artifact-review.mjs` when available. 9. Run manifest and HTML policy checks before handoff when the project has them. ## Output ### Artifact Family ### Decision Supported ### Canonical Source ### Generated Review Surface ### Owner Agent ### Evidence Links ### Fidelity Requirement ### Manifest Fields ### Readiness Gates ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [45ck](https://github.com/45ck) - **Source:** [45ck/skill-harness](https://github.com/45ck/skill-harness) - **License:** MIT - **Homepage:** https://github.com/45ck/skill-harness/releases/tag/v0.1.0 Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-45ck-skill-harness-visual-source-artifact - Seller: https://agentstack.voostack.com/s/45ck - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.