ARIS — Technology Adoption Intelligence

The System

Fixed DAG. Seven parallel branches. One personalised verdict.

Same input, same path, every time. A branch that fails lowers confidence, not the score — missing data is honest, not hidden. Click any pipeline stage to expand.

01 →

Decompose

Resolve tool, detect registry, validate queries.

An LLM agent resolves the tool name, detects the package registry, normalises the GitHub slug, and runs a validate_queries skill before any branch fires.

LLM agent

02 →

Seven Branches

All branches fire simultaneously.

Community · Production · Alternatives · Security · Downloads · GitHub Health · Stack Compatibility. Each has its own retry logic; a dropped branch lowers confidence, not the score.

LLM + Python

03 →

Score

PythonExec computes every dimension deterministically.

The custom PythonExec node runs deterministic Python inline. All six dimension scores use documented, named-constant formulas — no LLM touches a number. Repo verdict overrides the band (`ALREADY_USED` → `KEEP`).

Python only

04 →

Synthesize

LLM writes rationale grounded in your repo.

LLM receives computed scores, structured findings, and the repo profile — writes rationale, alternatives, caveats, and a per-dimension narrative that cites your specific stack. It cannot alter any figure.

LLM narrative

05 →

Deliver

HTML email + A4 PDF to your inbox.

Verdict badge, KPI tiles, repo scan block, engineer's-take pull-quote, six scored dimensions, rationale, alternatives, caveats. Readable in one glance. Auditable in full detail.

Python

Seven intelligence branches Hover a branch to see its data source

$ tavily

Community Sentiment

Tavily → LLM

Friction, enthusiasm, docs quality and setup complexity from real developer discourse.

LLM

$ prod-scan

Production Adoption

Tavily · GitHub · Stack Overflow

Enterprise case studies, stars, SO answers, package dependents. Boosted when your repo already uses the tool.

LLM + Python

$ alt-search

Alternatives

Tavily → LLM

Ranked alternatives and migration stories shaped by your use case.

LLM

$ osv-scan

Security Risk

OSV.dev → Python

Unpatched CVEs only. Age-penalised. Patched history excluded.

Python

$ dl-trend

Download Trajectory

PyPI · npm → LLM

Velocity: accelerating, stable, or declining. Trend context from Tavily.

LLM

$ gh-health

GitHub Health

GitHub API → Python

Commit velocity, bus factor, issue close-rate, release cadence.

Python

Repo-aware

$ aris scan_repo

Stack Compatibility

Your repo: manifests + README + tree + languages → Python + LLM

Parses requirements.txt, pyproject.toml, package.json, Pipfile, and setup.cfg. Detects your repo's archetype, identifies competing tools already declared, and produces a personalised engineer commentary displayed in the brief as a pull-quote. Verdict labels per-repo: ALREADY_USED · MIGRATION_REQUIRED · FIT · POOR_FIT.

Python + LLM

Inside your brief

What lands in your inbox.

Every brief has five layers: a verdict tile, a four-metric KPI strip, a terminal-style scan of your repo, an engineer's-take pull-quote written about your specific codebase, and six dimension scores with one-sentence narratives. Below is the actual format.

// VERDICT TILE

KEEP

Already in your stack

72.6 / 100

CONFIDENCE 95%

// QUICK SCAN

STACK FIT

SECURITY

100

MAINTENANCE

PRODUCTION

aris > scan_repo ● ok

description ─ A Django content management system focused on flexibility

archetype ─ web_service

key_libs ─ [django, requests, jest, webpack]

competitors ─ []

verdict ─ already_used

// ENGINEER'S TAKE

"We've already got djangorestframework pinned at >=3.15.1,<4.0, so the API layer is in place and works with the current Wagtail stack. The main thing to watch is the tight version ceiling; if we ever need DRF 4.x features, we'll have to coordinate a bump across the 41-package dependency graph. Overall, no adoption work is needed — just keep an eye on version compatibility as the ecosystem evolves."

// DIMENSION SCORES

stack_compatibilityStack Compatibility

Wagtail already declares DRF (>=3.15.1,<4.0); there are no competing API frameworks and migration effort is not applicable.

90/100

security_riskSecurity Risk

Three known CVEs, all patched; no unpatched critical or high-severity issues.

100/100

maintenance_healthMaintenance Health

Modest commit velocity (2.2/week) and high maintainer burnout risk.

48/100

Should you adopt that dependency?

Fixed DAG. Seven parallel branches. One personalised verdict.

What lands in your inbox.

Get your Adoption Brief