priorart-agent 0.1.0

Build-vs-borrow intelligence for agentic workflows - helps AI agents discover and evaluate open source packages

priorart is a deterministic package evaluation tool for build-vs-borrow decisions. Given a natural language task description and target language, it queries package registries directly, collects quantitative signals from GitHub and deps.dev, and produces a scored recommendation based on configurable, research-informed heuristics.

Research Inspiration

Noise-floor thresholds for registry metrics are informed by Koch et al. (MADWeb 2024), which quantified the weak, language-dependent correlation between GitHub stars and downstream adoption. Abandonment detection follows Coelho & Valente (ESEC/FSE 2017) on categorizing open-source project failure modes. Adoption saturation curves for committer diversity and reverse-dependency counts reference Borges & Valente (JSS 2018) and Zerouali et al. (ICSR 2018) on technical lag in dependency networks. Health dimensions are aligned with the CHAOSS Project metrics framework.

Pipeline

1. Taxonomy Mapping — Maps task descriptions to curated, language-specific registry search queries.
2. Registry Discovery — Fetches candidates from PyPI, npm, crates.io, or pkg.go.dev, ranked by download count.
3. Signal Collection — Enriches each candidate with GitHub repository metrics (stars, forks, MTTR, commit regularity) and deps.dev dependency health data.
4. Multidimensional Scoring — Computes weighted scores across reliability, adoption, versioning, activity regularity, and dependency health.
5. Decision Classification — Classifies packages as use_existing (≥75), evaluate (50–74), or build (<50).

Properties

• Registry-first discovery — Queries registries directly; does not rely on GitHub search.
• Latency — 50–200 ms cached, 3–5 s cold.
• Deterministic — Scoring is fully quantitative; no LLM-generated recommendations.
• Supply-chain checks — Identity verification (typosquatting), copyleft license detection, and dependency vulnerability flags.

Documentation

• SETUP.md: Installation, environment configuration, and MCP server setup.
• API.md: Comprehensive guide to the CLI, Python API, and MCP tool definitions.
• ARCHITECTURE.md: Deep dive into the scoring algorithms, data flow, and cache design.
• STYLE.md: Project coding standards and architectural invariants.
• TESTING.md: Guidelines for running unit and integration tests.
• AGENT_CONFIG.md: Specific protocols for AI agents using priorart in autonomous workflows.
• TAXONOMY.md: Guide for contributing new package categories and search terms.

License

See LICENSE file for details.