the bodies catch up to the renames, code-sync grows its second pass

@@ -1,10 +1,18 @@#!/bin/sh## sync.sh# code-sync.sh## Pull latest from origin for every git repo under ~/code/. Skips repos with# uncommitted changes. Uses --ff-only so divergent branches never get a silent# merge or rebase; they just warn and move on.# Two-pass sync for the ~/code/ workspace.## 1) Pull latest from origin for every git repo already under ~/code/. Skips#    repos with uncommitted changes or detached HEAD. Uses --ff-only so a#    divergent branch never gets a silent merge or rebase.## 2) Hit the public GitHub API for $GITHUB_USER (overshard) and clone any#    non-archived, non-fork, owned repos that don't exist locally yet, using#    the SSH key configured globally in ~/.ssh/config (no auth needed for the#    API call; this is one request per run, well under the 60/hr unauth#    rate limit).## Run after switching machines (e.g. desktop -> laptop) to catch up.#

@@ -12,6 +20,8 @@set -uCODE="$HOME/code"GITHUB_USER="overshard"if [ ! -d "$CODE" ]; then    echo "ERROR: $CODE does not exist" >&2    exit 1

@@ -20,6 +30,9 @@ fiok=0warn=0skip=0new=0echo "Pulling existing repos..."for path in "$CODE"/*/; do    [ -d "$path/.git" ] || continue

@@ -63,5 +76,42 @@ for path in "$CODE"/*/; do    fidonecd "$CODE"echo ""echo "Discovering repos for $GITHUB_USER..."api="https://api.github.com/users/$GITHUB_USER/repos?per_page=100&type=owner"json=$(curl -fsSL -H 'Accept: application/vnd.github+json' "$api" 2>/dev/null || true)if [ -z "$json" ]; then    printf "  [warn] GitHub API request failed\n"    warn=$((warn + 1))else    count=$(printf '%s' "$json" | jq 'length' 2>/dev/null || echo 0)    if [ "$count" -eq 100 ]; then        printf "  [warn] received exactly 100 repos; pagination may be needed\n"        warn=$((warn + 1))    fi    list=$(mktemp)    printf '%s' "$json" \        | jq -r '.[] | select(.archived == false and .fork == false) | "\(.name) \(.ssh_url)"' \        > "$list"    while IFS=' ' read -r repo url; do        [ -z "$repo" ] && continue        [ -d "$CODE/$repo" ] && continue        printf "  [new]  cloning %s\n" "$repo"        if git clone --quiet "$url" "$CODE/$repo" 2>/dev/null; then            new=$((new + 1))        else            printf "  [warn] clone failed for %s\n" "$repo"            warn=$((warn + 1))        fi    done < "$list"    rm -f "$list"fiecho ""printf "%d updated, %d warned, %d skipped\n" "$ok" "$warn" "$skip"printf "%d updated, %d cloned, %d warned, %d skipped\n" "$ok" "$new" "$warn" "$skip"

@@ -1,6 +1,6 @@#!/bin/sh## backup.sh# restic-backup.sh## Run manually to back up this container to Backblaze B2 via restic.# Initializes the repo on first run. Prunes per the retention policy

@@ -1,6 +1,6 @@#!/bin/sh## restore.sh# restic-restore.sh## Restore this container from the latest snapshot in Backblaze B2.# Existing contents of ~/.claude, ~/code, and ~/.ssh are moved aside to

@@ -1,6 +1,6 @@#!/bin/sh## status.sh# restic-status.sh## Show backup health for both restic repos (webdev and alpine) from a single# command. Both repos live in the same Backblaze B2 account, so the B2 creds