dhanoopbhaskar/discovery

Generates and maintains OKF v0.2-compliant knowledge bases documenting a codebase's modules, classes, functions, architecture, and configuration. Use when creating a knowledge base for a codebase, updating documentation after code changes, or exporting codebase knowledge in a portable, provenance-tracked format. Not for migration discovery or due-diligence mapping; use legacy-code-discovery for that.

Plans and runs a multi-phase, evidence-graded discovery program for an unfamiliar or legacy estate before a migration, replatform, rewrite, or major upgrade. Use when asked to produce a discovery dossier, estimate migration scope across a whole system, or coordinate reconnaissance, mapping, and risk/seam analysis over multiple sessions. Orchestrates repo-reconnaissance, system-cartography, and migration-seams and writes an OKF v0.2 knowledge bundle. Not for single-file changes, greenfield work, routine code navigation, or routine API documentation.

Turns a system map into migration decisions: scores risks, finds coupling hotspots and shared-data traps, identifies safe migration seams, assesses deployable readiness, compares migration options, and sequences candidate migration slices. Use as the risk-and-seams phase of legacy-code-discovery, or standalone when asked what could break, what must move together, where to start a migration, how to split a monolith or legacy service, or which risks threaten cutover. Prefer legacy-code-discovery when the request spans more than one phase or needs a durable bundle. Not for code review, architecture aesthetics, or selecting a destination stack.

Performs a fast, read-only structural survey of a repository, monorepo, or multi-repo estate before a migration, replatform, or due-diligence assessment. Use as the reconnaissance phase of legacy-code-discovery, or standalone when you need to inventory a codebase for migration scope: which stacks and versions it uses, how it builds, tests, and deploys, which deployable units exist, what should be excluded, how to split a large repository into workable analysis shards, or how to emit a project profile. Prefer legacy-code-discovery when the request spans more than one phase or needs a durable bundle. Produces a repository registry, stack map, coverage ledger, and, when requested, .sdlc/profile.json. Not for everyday code exploration, feature work, or behavior analysis.

Maps the internal and external structure of a legacy system for a migration or due-diligence assessment: components, dependency edges, entry points, end-to-end flows, integration contracts, data ownership, and runtime topology. Use as the cartography phase of legacy-code-discovery, or standalone during a migration or due-diligence assessment when asked what depends on what, where requests and events go, or what data is shared. Prefer legacy-code-discovery when the request spans more than one phase or needs a durable bundle. Works from source, build metadata, and contracts without assuming a tech stack. Not for routine code navigation or single-feature analysis.