kb-load

name	kb-load
description	Proactively warm up context from the Ecosystem KB (prograph-vault) when starting work on a known area — interfaces, cross-cutting rules, contracts, or established patterns for the current project. Loads the relevant bundle at once (applicable rules + this project's facts/interfaces + related contracts + recent journal). Use at the start of a work area; for a single targeted question use kb-search instead.
allowed-tools	Bash, Read, Grep, Glob

kb-load — warm up work-area context from the KB

Front-loads the set of KB knowledge relevant to what you are about to work on, so the rules, interfaces and patterns are in context before you start. This is the proactive bulk counterpart to kb-search (which answers one ad-hoc question).

Step 0 — Locate the KB and the project

_kb() { local d; d="$(pwd -P)"; while [ "$d" != "/" ]; do
    [ -d "$d/prograph-vault/authored" ] && { printf '%s\n' "$d/prograph-vault"; return 0; }
    [ -d "$d/authored" ] && [ -f "$d/CLAUDE.md" ] && [ "$(basename "$d")" = prograph-vault ] && { printf '%s\n' "$d"; return 0; }
    d="$(dirname "$d")"; done; return 1; }
KB_ROOT="${KB_ROOT:-$(_kb)}"
[ -z "$KB_ROOT" ] && { echo "ℹ️  prograph-vault not found — continuing without the KB."; exit 0; }
source "$KB_ROOT/authored/skills/kb-utils/kb-env.sh"
PROJECT="$(kb_project "${1:-}" || true)"

PROJECT = explicit argument, else the current ecosystem repo. TOPIC = the rest of the argument or the conversation context (e.g. gui, tui, contracts, code-style, an interface name).

Step 1 — Load the relevant bundle (read these into context)

Pull the pieces that match the project and topic — do not read the whole KB:

# Cross-cutting rules (all, or the ones matching the topic: gui/tui/code-style/libraries)
ls "$KB_ROOT/authored/rules/"
# read the applicable ones, e.g.:
#   Read "$KB_ROOT/authored/rules/<topic>.md"

# This project's facts / interfaces (written by prograph)
ls "$KB_ROOT/derived/projects/${PROJECT}"* 2>/dev/null

# Contracts this project produces or consumes (snapshots — authority is in the producing repo)
kb_grep "${PROJECT}" "$KB_ROOT/derived/contracts" 2>/dev/null

# Recent activity for this project
tail -n 60 "$KB_ROOT/derived/journal/${PROJECT}/journal.md" 2>/dev/null

# Templates / scaffolds, when present
ls "$KB_ROOT/authored/templates/" 2>/dev/null

Use Read/Grep to open the files the listings surface. Prefer the topic-relevant subset.

Step 2 — Summarize what was loaded

• Which files were pulled in (paths).
• Key rules / interfaces / patterns that apply here (3–6 bullets).
• Anything expected but missing — if the area is new to the KB, suggest /kb-save to record work as it happens.

Reading and search are unrestricted across the whole KB. This skill does not write.