Methodology
Sector Radar Method
Denominator: In-cluster, N=22. Source: accepted SC-024 report, centroid, visual data, and calibration artifacts. Reader meaning: this page explains how to read the public report and what the numbers do and do not claim.
Principle
Sector Radar maps convergence and differentiation in public software ecosystems. It does not rank maintainers, infer authorship, assign universal product quality, or publish private strategy verdicts.
Presentation may summarize analysis, but it must not create analysis. Scores, roles, health metrics, centroid traits, and graph relationships stay grounded in source artifacts.
Denominator
The final denominator is In-cluster, N=22. The corpus starts from SC-021 source-verified accepted candidates. SC-024 refreshed GitHub metadata and harvested local evidence packets for all 22 before centroid construction and scoring.
No accepted candidate was dropped during SC-024, and no Issue #19 metadata-only candidate was added. Issue #19 remains useful scout context, but it is not in-cluster truth for this public report.
| Metric family | Denominator / sample | Source | Reader meaning |
|---|---|---|---|
| Centroid traits | In-cluster, N=22 | Centroid artifact | Shows the sector default from accepted candidates only. |
| Scores and roles | In-cluster, N=22 | Calibration table | Shows sector position at the snapshot date. |
| Health metrics | In-cluster, N=22 unless a row names a subset | Visual data and calibration table | Summarizes score matrix patterns without changing row-level evidence. |
| Protocol/runtime surfaces | In-cluster, N=22 | Visual data Table 8 | Shows boundary-shaping surfaces that explain differentiation. |
Corpus Boundary
Repos must be public, fresh, non-archived, package-backed VS Code or VS Code-compatible extension surfaces whose headline behavior is AI coding assistant, coding agent, or coding-agent client workflow.
Editor forks, generic extension samples, closed-source products, standalone CLIs, MCP servers without agent-client extension surfaces, ACP runtimes without an extension client surface, and unrelated AI utilities are excluded.
Prompt/apply, MCP-client, and ACP-client repos remain visible where source evidence supports a VS Code coding-agent client surface. They do not erase the distinction between an agent runtime and a protocol client.
Scoring Discipline
Scores are evidence summaries, not universal quality judgments. Role labels describe sector position at the snapshot date.
| Dimension | Question | Reader meaning |
|---|---|---|
| CM | How closely does this repo match the centroid? | Closeness to the default VS Code agent-extension pattern. |
| DF | What does it offer beyond the centroid? | Observable differentiation beyond the default. |
| HE | How much sustained engineering is visible? | Reviewable depth in code, package structure, docs, tests, and maintenance. |
| AD | How much adoption signal exists? | Public reach through stars, forks, issues, and ecosystem visibility. |
| CI | Do claims match source evidence? | Whether README/manifest claims trace to package and code evidence. |
Public Language
Public role labels are Category Leader, Differentiated Niche, Strong Default, Baseline Implementation, Inactive Scaffold, Adjacent Infrastructure, Reference Implementation, and Edge Case. This report uses only the labels that appear in the accepted VS Code corpus.
The report describes observable sector position. It avoids private action-label vocabulary and negative maintainer characterization. "Baseline Implementation" means close to the default with limited independent signal at the snapshot date; it is not a judgment about maintainers.
Evidence Graph Interpretation
The evidence graph connects repositories through inspectable evidence fields: shared centroid traits, protocol/runtime surfaces, workflow themes, lineage/client-family notes, and adjacent boundary signals.
Edges and proximity do not imply causation, authorship, derivation, dependency, quality, or endorsement. Layout coordinates are visual presentation; edge reasons and evidence references are the auditable fields.
Open the full evidence graph or return to the embedded report view.
Limitations
| Limitation | How this report handles it | Reader meaning |
|---|---|---|
| Time-sensitive source facts | Evidence snapshot is visible and corrections are invited. | Stars, archive state, package structure, and README claims can change after publication. |
| Runtime behavior not exhaustively executed | SC-024 harvested local repos and metadata but did not run every extension, install every package, or execute every test. | Source/package evidence is stronger than metadata-only claims, but not a full runtime review. |
| Protocol boundary pressure | MCP and ACP surfaces are separated and labeled. | Different client/runtime relationships should not be merged into one protocol bucket. |
| Lineage risk | Lineage-bearing rows carry explicit notes instead of hidden overrides. | Lineage affects interpretation but does not bypass source evidence. |
| Future bridge relevance | The report records contribution evidence but does not publish a Super Cluster graph. | Bridge framing remains bounded to accepted SC-024 findings. |
Source Artifacts
Source map for the methodology page. The public methodology page explains how to read the report; canonical scoring definitions remain in Manual-audit/AUDIT-TEMPLATE-SECTOR-RADAR-v1.md.
| Published artifact | Source role | Visible metric families |
|---|---|---|
| Reports/SECTOR-RADAR-VS-CODE-AI-CODING-AGENT-EXTENSIONS.md | Public narrative source | Methodology summary, corpus boundary, limitations, findings. |
| Manual-audit/CENTROID-VS-CODE-AI-CODING-AGENT-EXTENSIONS.md | Centroid source | Sector definition, inclusion/exclusion rules, denominator, trait threshold. |
| Manual-audit/CALIBRATION-TABLE-RADAR-VS-CODE-AI-CODING-AGENT-EXTENSIONS.md | Calibration source | Score dimensions, public role mapping, decision rules, boundary register. |
| Reports/VISUAL-DATA-VS-CODE-AI-CODING-AGENT-EXTENSIONS.md | Data source | Metric denominators, health table, protocol/runtime distribution. |
| Operations/SECTOR-RADAR-DISPLAY-STANDARD.md | Display contract | Denominator, Source, Reader meaning, public vocabulary, evidence graph, correction path. |
| Operations/PROTOCOLS.md | Public-site and sprint protocol | Release checks and standing-standard requirements. |
Corrections
Found a stale repo, wrong source trace, or boundary classification issue? Send a correction to hello@diversum.dev with the repo name, section, disputed claim or score, and source evidence.