What problem does it solve? Before designing a software system, teams often rebuild what already exists because they never systematically searched the open-source world. This Skill runs a reproducible prior-art survey — deriving a typed keyword map, executing search angles across dozens of sources, deep-reading candidate repositories, and synthesizing findings — so design decisions rest on evidence rather than guesswork. ## Core Features & Use Cases - Keyword-map derivation: Distills a scope description into typed search vocabulary (domain, capability, technique, ecosystem anchor, community, competitor groups) with provenance-stamped expansions, negative terms, and a community-vocabulary probe. - Nine search angles: Executes one discovery mechanism per run — host metadata search, curated catalogs, package registries and dependents graphs, code-content search, alternative directories, community mining, academic code, model hubs, and platform marketplaces — with machine-checkable coverage cells proving every group-by-source pair was worked. - Extraction and synthesis: Deep-reads one candidate repo into a 10-section extraction with verdict, 0-10 score, SPDX license, and purl dependencies, then aggregates the extract set into an eleven-section report plus a machine-readable borrow-index. - Schema-validated outputs: Every artifact (keyword map, search output, extraction, borrow-index) is gated by JSON Schemas and a deterministic validator script that must exit 0 before handoff. - Use Case: Before building a backtesting engine, run the survey to derive the keyword map, execute the registry and community angles, extract the top five candidate repos, and produce a borrow-vs-build report showing which subsystems have production-grade prior art. ## Quick Start Ask the agent to run a code prior art survey for your software idea, starting with deriving the keyword map from your scope description.
