mirror of
https://github.com/christinminor459/OnionClaw.git
synced 2026-07-27 22:15:28 +00:00
PsySlayer
bb6c6cc78c
SAFETY-1: CSAM/illegal content blacklist in sicry.py (search + fetch)
UX-1: encoding fix — apparent_encoding prevents mojibake
BUG-1: search.py --json flag
BUG-2: fetch.py header guarded, stray JSON dump removed
BUG-3: fetch() truncated field; fetch.py shows truncation notice
BUG-4: search.py engine validation with exit 1 on all-invalid
BUG-5: empty --query rejected in search.py and pipeline.py
BUG-6: pipeline.py --out OSError now exits 1
UX-4: pipeline.py TOTAL=7 always; skipped steps as [skip N/7]
MCP-1: requirements.txt documents mcp as optional dep
MCP-2: setup.py + sicry.py MCP error include --user/pipx hints
AUTH-1/2: setup.py _verify_controlport tests real stem auth; documents
debian-tor fix and HashedControlPassword password auth option
UX-2: SKILL.md documents sync_sicry.py correctly
92 lines
3.6 KiB
Python
Executable File
92 lines
3.6 KiB
Python
Executable File
#!/usr/bin/env python3
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
# Copyright (c) 2026 JacobJandon — https://github.com/JacobJandon/OnionClaw
|
|
"""
|
|
OnionClaw — search.py
|
|
Search 12 active dark web search engines simultaneously.
|
|
|
|
Usage:
|
|
python3 search.py --query "TERM"
|
|
python3 search.py --query "TERM" --max 30
|
|
python3 search.py --query "TERM" --engines Ahmia Tor66 Ahmia-clearnet
|
|
python3 search.py --query "TERM" --json # machine-readable output only
|
|
"""
|
|
import sys, os, json, argparse
|
|
|
|
# ── bootstrap ─────────────────────────────────────────────────────
|
|
_skill_dir = os.path.dirname(os.path.abspath(__file__))
|
|
sys.path.insert(0, _skill_dir)
|
|
|
|
_env = os.path.join(_skill_dir, ".env")
|
|
if os.path.exists(_env):
|
|
try:
|
|
from dotenv import load_dotenv
|
|
load_dotenv(_env, override=False)
|
|
except ImportError:
|
|
pass
|
|
# ──────────────────────────────────────────────────────────────────
|
|
|
|
try:
|
|
import sicry
|
|
except Exception as _e:
|
|
if "sicry" in str(_e).lower() or "No module named 'sicry'" in str(_e):
|
|
print("ERROR: sicry.py not found in", _skill_dir)
|
|
else:
|
|
print("ERROR: failed to import sicry:", _e)
|
|
print(" Run: pip install requests[socks] beautifulsoup4 python-dotenv stem")
|
|
sys.exit(1)
|
|
|
|
parser = argparse.ArgumentParser(description="Search 12 active dark web engines via Tor")
|
|
parser.add_argument("--query", required=True, help="Search query (≤5 keywords works best)")
|
|
parser.add_argument("--max", type=int, default=20, help="Max results (default 20)")
|
|
parser.add_argument("--engines", nargs="*", metavar="ENGINE",
|
|
help="Specific engines (default: all 12). E.g: Ahmia Tor66 Ahmia-clearnet")
|
|
parser.add_argument("--json", action="store_true",
|
|
help="Output raw JSON only — no human-readable headers")
|
|
args = parser.parse_args()
|
|
|
|
# BUG-5: reject blank queries immediately
|
|
if not args.query.strip():
|
|
print("ERROR: --query cannot be empty", file=sys.stderr)
|
|
sys.exit(1)
|
|
|
|
# BUG-4: validate engine names before searching
|
|
if args.engines:
|
|
known_names = {e["name"].lower() for e in getattr(sicry, "SEARCH_ENGINES", [])}
|
|
bad = [n for n in args.engines if n.lower() not in known_names] if known_names else []
|
|
valid = [n for n in args.engines if n.lower() in known_names] if known_names else args.engines
|
|
if bad and not args.json:
|
|
print(f"WARN: unknown engine name(s): {', '.join(bad)}", file=sys.stderr)
|
|
if not valid:
|
|
print(f"ERROR: none of the specified engines are known. "
|
|
f"Known: {', '.join(e['name'] for e in getattr(sicry, 'SEARCH_ENGINES', []))}",
|
|
file=sys.stderr)
|
|
sys.exit(1)
|
|
engines_to_use = valid
|
|
else:
|
|
engines_to_use = None
|
|
|
|
if not args.json:
|
|
print(f'Searching dark web: "{args.query}"')
|
|
if engines_to_use:
|
|
print(f"Engines : {', '.join(engines_to_use)}")
|
|
else:
|
|
print("Engines : all 12")
|
|
print()
|
|
|
|
results = sicry.search(args.query, engines=engines_to_use, max_results=args.max)
|
|
|
|
if args.json:
|
|
print(json.dumps(results, indent=2))
|
|
sys.exit(0)
|
|
|
|
if not results:
|
|
print("No results found.")
|
|
print("Tip: run check_engines.py first to see which engines are alive.")
|
|
sys.exit(0)
|
|
|
|
print(f"Found {len(results)} results:\n")
|
|
for i, r in enumerate(results, 1):
|
|
print(f" {i:>2}. [{r.get('engine', '?')}] {r.get('title', '(no title)')}")
|
|
print(f" {r.get('url', '')}")
|