Repo Intelligence Enrichment (repo-intel/enrich.js)

Collects README, manifest and hotspot data, builds prompts for summarizer and weighter agents, and integrates their outputs into repository intelligence.

// Source: https://github.com/mperkins0155/agentsys | License: MIT | Copyright 2026 Avi Fenesh
// Provided via justclaw.it.com; keep this notice with the code.
/**
 * Post-init enrichment helpers.
 *
 * The repo-intel skill spawns two Haiku-backed Task subagents after
 * the deterministic init/update pass:
 *
 * 1. `repo-intel-summarizer` reads README + manifests + hotspot
 *    headers and writes a 3-depth narrative summary.
 * 2. `repo-intel-weighter` writes 1-2 sentence descriptors for the
 *    top-N most-active files, used by `find` to add semantic recall.
 *
 * The orchestration calls into this module to gather the agents'
 * inputs, parse their JSON outputs (which arrive between marker
 * blocks because subagent stdout is otherwise free-form), and pipe
 * the result back through `re

... (truncated -- full source via MCP)