#!/bin/bash
# webhook-audit-alert.sh - Daily webhook audit summary
#
# Analyzes the OpenClaw webhook event logs (JSONL) for the previous day:
#   - Total events, by result type
#   - Blocked/rejected events (anomalies)
#   - Signature failures
#   - Untrusted sender attempts
#
# Posts summary to a Mattermost channel and creates Gitea issues for anomalies.
#
# Daily cron:
#   5 0 * * * /opt/webhook-security/scripts/webhook-audit-alert.sh >> /var/log/webhook-audit.log 2>&1
#
# Prerequisites:
#   - jq installed
#   - MATTERMOST_BOT_TOKEN env var set
#   - Gitea admin token file at GITEA_TOKEN_FILE path
# ------------------------------------------------------------------

set -euo pipefail

# ========================= CONFIGURATION =========================
# Customize these values for your environment

LOGS_DIR="/var/lib/openclaw/hooks/logs"                    # OpenClaw webhook event log directory
MATTERMOST_URL="@@MATTERMOST_URL@@"                      # e.g., https://mattermost.example.com/api/v4
MATTERMOST_TOKEN="${MATTERMOST_BOT_TOKEN:?Set MATTERMOST_BOT_TOKEN env var}"
MATTERMOST_CHANNEL="@@MATTERMOST_CHANNEL_ID@@"           # Target channel for audit summaries
GITEA_API="@@GITEA_API_BASE@@/api/v1"              # e.g., https://git.example.com/api/v1
GITEA_TOKEN_FILE="/etc/nginx/gitea-admin-token"
GITEA_REPO="@@GITEA_REPO@@"                              # e.g., myorg/webhook-security
LOG_TAG="webhook-audit"

# =================================================================

# Date for analysis (yesterday)
AUDIT_DATE=$(date -u -d "yesterday" '+%Y-%m-%d' 2>/dev/null || date -u -v-1d '+%Y-%m-%d')
MONTH=$(echo "$AUDIT_DATE" | cut -d- -f1-2)
LOG_FILE="$LOGS_DIR/webhook-events-${MONTH}.jsonl"

log() {
    echo "[$(date -u '+%Y-%m-%d %H:%M:%S UTC')] $1"
}

post_mattermost() {
    local MSG="$1"
    local ROOT_ID="${2:-}"
    local PAYLOAD
    
    if [ -n "$ROOT_ID" ]; then
        PAYLOAD=$(jq -n --arg ch "$MATTERMOST_CHANNEL" --arg msg "$MSG" --arg rid "$ROOT_ID" \
            '{channel_id: $ch, message: $msg, root_id: $rid}')
    else
        PAYLOAD=$(jq -n --arg ch "$MATTERMOST_CHANNEL" --arg msg "$MSG" \
            '{channel_id: $ch, message: $msg}')
    fi
    
    RESPONSE=$(curl -sf -X POST "$MATTERMOST_URL/posts" \
        -H "Authorization: Bearer $MATTERMOST_TOKEN" \
        -H "Content-Type: application/json" \
        -d "$PAYLOAD" 2>/dev/null || echo '{"id":""}')
    
    echo "$RESPONSE" | jq -r '.id' 2>/dev/null || echo ""
}

create_gitea_issue() {
    local TITLE="$1"
    local BODY="$2"
    
    if [ ! -f "$GITEA_TOKEN_FILE" ]; then
        log "WARNING: No Gitea token file, skipping issue creation"
        return
    fi
    
    local TOKEN=$(cat "$GITEA_TOKEN_FILE" | tr -d '[:space:]')
    
    curl -sf -X POST "$GITEA_API/repos/$GITEA_REPO/issues" \
        -H "Authorization: token $TOKEN" \
        -H "Content-Type: application/json" \
        -d "$(jq -n --arg t "$TITLE" --arg b "$BODY" '{title: $t, body: $b, labels: []}' )" \
        > /dev/null 2>&1 || log "WARNING: Failed to create Gitea issue"
}

log "Starting audit for $AUDIT_DATE"

# Check if log file exists
if [ ! -f "$LOG_FILE" ]; then
    log "No log file found: $LOG_FILE"
    exit 0
fi

# Filter to yesterday's entries
YESTERDAY_ENTRIES=$(grep "\"$AUDIT_DATE" "$LOG_FILE" || true)

if [ -z "$YESTERDAY_ENTRIES" ]; then
    log "No webhook events on $AUDIT_DATE"
    exit 0
fi

# Count totals
TOTAL=$(echo "$YESTERDAY_ENTRIES" | wc -l)
ROUTED=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"routed"' || echo 0)
SKIPPED_LOOP=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"skipped_loop"' || echo 0)
SKIPPED_DUP=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"skipped_dup"' || echo 0)
SKIPPED_UNSUPPORTED=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"skipped_unsupported"' || echo 0)
REJECTED_NO_SIG=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"rejected_no_sig"' || echo 0)
REJECTED_NO_SENDER=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"rejected_no_sender"' || echo 0)
PERM_TRUSTED=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"perm_trusted"' || echo 0)
PERM_UNTRUSTED=$(echo "$YESTERDAY_ENTRIES" | grep -c '"result":"perm_untrusted"' || echo 0)

# Check for anomalies
ANOMALIES=0
ANOMALY_DETAILS=""

if [ "$REJECTED_NO_SIG" -gt 0 ]; then
    ANOMALIES=$((ANOMALIES + REJECTED_NO_SIG))
    ANOMALY_DETAILS="${ANOMALY_DETAILS}\n- $REJECTED_NO_SIG events with missing signature"
fi

if [ "$REJECTED_NO_SENDER" -gt 0 ]; then
    ANOMALIES=$((ANOMALIES + REJECTED_NO_SENDER))
    ANOMALY_DETAILS="${ANOMALY_DETAILS}\n- $REJECTED_NO_SENDER events with missing sender"
fi

if [ "$PERM_UNTRUSTED" -gt 0 ]; then
    # Get unique untrusted senders
    UNTRUSTED_SENDERS=$(echo "$YESTERDAY_ENTRIES" | grep '"result":"perm_untrusted"' | jq -r '.sender' 2>/dev/null | sort -u | tr '\n' ', ' | sed 's/,$//')
    ANOMALY_DETAILS="${ANOMALY_DETAILS}\n- $PERM_UNTRUSTED events from untrusted senders: $UNTRUSTED_SENDERS"
fi

# Build summary message
SUMMARY="### Webhook Audit Summary: $AUDIT_DATE\n\n"
SUMMARY="${SUMMARY}| Metric | Count |\n|---|---|\n"
SUMMARY="${SUMMARY}| Total events | $TOTAL |\n"
SUMMARY="${SUMMARY}| Routed (delivered) | $ROUTED |\n"
SUMMARY="${SUMMARY}| Skipped (loop prevention) | $SKIPPED_LOOP |\n"
SUMMARY="${SUMMARY}| Skipped (duplicate) | $SKIPPED_DUP |\n"
SUMMARY="${SUMMARY}| Skipped (unsupported event) | $SKIPPED_UNSUPPORTED |\n"
SUMMARY="${SUMMARY}| Rejected (no signature) | $REJECTED_NO_SIG |\n"
SUMMARY="${SUMMARY}| Rejected (no sender) | $REJECTED_NO_SENDER |\n"
SUMMARY="${SUMMARY}| Permission: trusted | $PERM_TRUSTED |\n"
SUMMARY="${SUMMARY}| Permission: untrusted | $PERM_UNTRUSTED |\n"

if [ "$ANOMALIES" -gt 0 ]; then
    SUMMARY="${SUMMARY}\n**ANOMALIES DETECTED ($ANOMALIES):**$(echo -e "$ANOMALY_DETAILS")\n"
fi

# Post to Mattermost
log "Posting summary to Mattermost"
ROOT_POST_ID=$(post_mattermost "$(echo -e "$SUMMARY")")

if [ -n "$ROOT_POST_ID" ] && [ "$ROOT_POST_ID" != "null" ] && [ "$ROOT_POST_ID" != "" ]; then
    log "Posted to Mattermost, post ID: $ROOT_POST_ID"
else
    log "WARNING: Failed to post to Mattermost"
fi

# Create Gitea issue for anomalies
if [ "$ANOMALIES" -gt 0 ]; then
    log "Creating Gitea issue for $ANOMALIES anomalies"
    ISSUE_TITLE="[AUDIT] $ANOMALIES anomalies detected on $AUDIT_DATE"
    ISSUE_BODY="## Webhook Audit Anomalies: $AUDIT_DATE\n\n**$ANOMALIES anomalous events detected:**\n$(echo -e "$ANOMALY_DETAILS")\n\n### Action Required\nReview the webhook event log at \`$LOG_FILE\` for entries from $AUDIT_DATE.\n\nFilter command:\n\`\`\`bash\ngrep \"$AUDIT_DATE\" $LOG_FILE | jq 'select(.result | test(\"rejected|untrusted\"))'\n\`\`\`"
    create_gitea_issue "$ISSUE_TITLE" "$(echo -e "$ISSUE_BODY")"
fi

log "Audit complete. Total: $TOTAL, Anomalies: $ANOMALIES"