dondai1234/master-fetch

Review signals: 1 connection review. 3 static matches retained as context.

Reviewed as expected for a test run: attempted writes that failed, so nothing was written.

Repository processes attempted these non-loopback connections; proxies can hide the final hostname. 9 distinct locations. 17 attributed processes.

2 hardcoded secret matches retained as context; full list in the evidence.

OSV scanned supported lockfiles; unpinned dependencies and unsupported formats are outside this check.

A short run cannot establish that code is safe.

Connections show IP addresses; proxies can hide the final hostname.

Tracing covers dispatched commands and their children, not the whole OS; guest root can interfere with tracing.

Full safety evidence →

• smart_fetch fetches a URL over HTTP and returns its content: Real input produced the expected result: This domain is for use in documentation examples

  Source: Fetch any URL. HTTP first, auto-escalates to the anti-detect browser if blocked.

  Input: {"tool": "mcp_smart_fetch", "arguments": {"url": "https://example.com/", "cache_ttl": 0}}

  Expected: page content returned

  Command: <guest-path> tools/call {"tool": "mcp_smart_fetch", "arguments": {"url": "https://example.com/", "cache_ttl": 0}}

  Observed: …This domain is for use in documentation examples without needing permission. This is not a service; avoid relying on it for testing and monitoring purposes.\"],\"url\":\"https://example.com/\",\"cached\":false,\"fetcher_used\":\"http\",\"extracted_type\":\"markdown\",\"session_id\":\"\",\"duration_ms\":649.630069732666,\"error\":\"\",\"content_type\":\"text/html; charset=utf-8\",\"total_size_bytes\":577,\"total_extracted_chars\":156,\"is_truncated\":false,\"next_offset\":0,\"escalation_path\":\"direct:http\",\"retry_count\":0,\"summary\":\"200 OK · 577B markdown · http\",\"content_ok\":true,\"

Examples

• Repository's documented offline test suite (CI command) runs: command exited with code 1: te-packages/pytest_asyncio/plugin.py:300: PytestDeprecationWarning: The configuration option "asyncio_default_fixture_loop_scope" is unset. The event loop scope for asynchronous fixtures will default to the "fixture" caching scope. Future versions of pytest-asyncio will default the loop scope for asynchronous fixtures to "function" scope. Set the default fixture loop scope explicitly in order to avoid unexpected behavior in the future. Valid fixture loop scopes are: "function", "class", "module", "package", "session" warnings.warn(PytestDeprecationWarning(_DEFAULT_FIXTURE_LOOP_SCOPE_UNSET))

  Source: .github/workflows/ci.yml, Run tests step

  Input: python -m pytest tests/ -v --tb=short

  Expected: 762 tests pass; one date-dependent test fails on the VM clock (2026-10-28)

  Command: <guest-path> -m pytest tests/ -v --tb=short

  Observed: te-packages/pytest_asyncio/plugin.py:300: PytestDeprecationWarning: The configuration option "asyncio_default_fixture_loop_scope" is unset. The event loop scope for asynchronous fixtures will default to the "fixture" caching scope. Future versions of pytest-asyncio will default the loop scope for asynchronous fixtures to "function" scope. Set the default fixture loop scope explicitly in order to avoid unexpected behavior in the future. Valid fixture loop scopes are: "function", "class", "module", "package", "session" warnings.warn(PytestDeprecationWarning(_DEFAULT_FIXTURE_LOOP_SCOPE_UNSET))

Local or hosted: local

Run the hound MCP server over stdio and call mcp_smart_fetch to fetch https://example.com/ (HTTP tier), returning page content

cd <guest-path> && ls -la && cat README.md | head -c 6000 && python3 --version && echo "CDP=$CDP_URL" && curl -s -m 10 -i $CDP_URL/json/version | head -20
cd <guest-path> && grep -n "^## \|^### " README.md | head -80
cd <guest-path> && awk '/^## .*Install/{f=1} /^## .*(The 6 tools|Tools)/{if(f)exit} f' README.md | head -c 7000
cd <guest-path> && sed -n 81,100p README.md; sed -n 146,158p README.md
cd <guest-path> && grep -n "requires-python\|^name\|^version" pyproject.toml; python3 -m venv --help >/dev/null 2>&1 && echo venv_ok || echo no_venv; ls .coldpass
mkdir -p <guest-path> && python3 -m venv <guest-path> && <guest-path> install -q --upgrade pip 2>&1 | tail -3; <guest-path> install "hound-mcp[all]" 2>&1 | tail -5
cd /tmp && <guest-path> uninstall -y -q hound-mcp && <guest-path> install -q "<guest-path>" 2>&1 | tail -5; <guest-path> show hound-mcp | head -2; cd <guest-path> && git status 2>&1 | head -3; ls -a
cd /tmp && H=<guest-path>; $H -v 2>&1 | head -5; echo "rc=$?"; $H --help 2>&1 | head -40
cd /tmp && printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"t","version":"1"}}}' '{"jsonrpc":"2.0","method":"notifications/initialized"}' '{"jsonrpc":"2.0","id":2,"method":"tools/list"}' | timeout 60 <guest-path> 2><guest-path> | python3 -c 'import sys,json
for l in sys.stdin:
    d=json.loads(l)
    if d.get("id")==2:
        for t in d["result"]["tools"]:
            print(t["name"], list(t.get("inputSchema",{}).get("properties",{}).keys()), t.get("inputSchema",{}).get("required"))
'; echo rc=$?; head -c 300 <guest-path>
cd /tmp && cat <guest-path> | tail -25
cd <guest-path> && grep -n -i "mcp" pyproject.toml; <guest-path> show mcp | head -2; ls <guest-path> && grep -n "pip\|mcp" .github/workflows/*.yml | head -20
cd /tmp && <guest-path> install -q "mcp>=1.27.0,<2" 2>&1 | tail -3; <guest-path> show mcp | head -2; printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"t","version":"1"}}}' '{"jsonrpc":"2.0","method":"notifications/initialized"}' '{"jsonrpc":"2.0","id":2,"method":"tools/list"}' | timeout 60 <guest-path> 2><guest-path> > <guest-path>; echo rc=$?; python3 -I -c 'import json
for l in open("<guest-path>"):
    d=json.loads(l)
    if d.get("id")==2:
        for t in d["result"]["tools"]:
            print(t["name"], list(t.get("inputSchema",{}).get("properties",{}).keys()), t.get("inputSchema",{}).get("required"))
'; tail -3 <guest-path>
cd /tmp && grep -o '"name": *"[^"]*"' <guest-path> | sort -u | head; printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"t","version":"1"}}}' '{"jsonrpc":"2.0","method":"notifications/initialized"}' '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"version","arguments":{}}}' '{"jsonrpc":"2.0","id":4,"method":"tools/call","params":{"name":"mcp_smart_fetch","arguments":{"url":"https://example.com/","cache_ttl":0}}}' | timeout 120 <guest-path> 2><guest-path> | head -c 2500; echo; echo rc=$?
cat <guest-path> | tail -30
cd /tmp && ( printf '%s\n' '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-03-26","capabilities":{},"clientInfo":{"name":"t","version":"1"}}}' '{"jsonrpc":"2.0","method":"notifications/initialized"}' '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"version","arguments":{}}}'; sleep 20; printf '%s\n' '{"jsonrpc":"2.0","id":4,"method":"tools/call","params":{"name":"mcp_smart_fetch","arguments":{"url":"https://example.com/","cache_ttl":0}}}'; sleep 60 ) | timeout 110 <guest-path> 2><guest-path> | cut -c1-1500; echo rc=$?; tail -c 800 <guest-path>
cd <guest-path> && mkdir -p .coldpass && cat > .coldpass/probe.json <<'EOF'
{"kind":"mcp","argv":["<guest-path>"],"cwd":"<guest-path>","tool":"mcp_smart_fetch","arguments":{"url":"https://example.com/","cache_ttl":0},"expected_text":"This domain is for use in documentation examples"}
EOF
cat > .coldpass/claims.json <<'EOF'
[{"kind":"claim","claim":"smart_fetch fetches a URL over HTTP and returns its content","source":"Fetch any URL. HTTP first, auto-escalates to the anti-detect browser if blocked.","input":"https://example.com/","expected":"page content returned","probe_kind":"mcp","argv":["<guest-path>"],"tool":"mcp_smart_fetch","arguments":{"url":"https://example.com/","cache_ttl":0},"expected_text":"This domain is for use in documentation examples","cwd":"<guest-path>"}]
EOF
python3 -I -m json.tool .coldpass/probe.json >/dev/null && python3 -I -m json.tool .coldpass/claims.json >/dev/null && echo ok; grep -c "Fetch any URL. HTTP first, auto-escalates to the anti-detect browser if blocked." README.md
cd <guest-path> && grep -rn -i "cron\|schedul\|--once\|webhook" src README.md --include=*.py --include=*.md | head -10; echo ---; sed -n 30,60p .github/workflows/ci.yml
cd <guest-path> && sed -n '/optional-dependencies/,/^\[tool/p' pyproject.toml | head -30
cd /tmp && <guest-path> install -q "pytest>=8.0" "pytest-asyncio>=0.24" "pytest-mock>=3.14" 2>&1 | tail -3; cd <guest-path> && timeout 900 <guest-path> -m pytest tests/ -q --tb=line -p no:cacheprovider 2>&1 | tail -15
cd <guest-path> && sed -n 270,290p tests/test_envelope.py
cd <guest-path> && cat > .coldpass/examples.json <<'EOF'
[{"kind":"example","claim":"Repository's documented offline test suite (CI command) runs","source":".github/workflows/ci.yml, Run tests step","input":"python -m pytest tests/ -v --tb=short","expected":"762 tests pass; one date-dependent test fails on the VM clock (2026-10-28)","argv":["<guest-path>","-m","pytest","tests/","-v","--tb=short"],"cwd":"<guest-path>","env":{},"expected_text":"762 passed","expected_returncode":1,"expected_exit_meaning":"one test failure in tests/test_envelope.py (test-owned date assumption)"}]
EOF
cat > .coldpass/cycles.json <<'EOF'
{"detected":false,"checks":[],"reason":"Inspected README.md and src/master_fetch (cli.py, server.py): Hound is a stdio/streamable-HTTP MCP server with no cron, scheduler, --once/--now, loop, or webhook trigger. Only CI schedules and cache TTLs exist, which are not product cycles."}
EOF
python3 -I -m json.tool .coldpass/examples.json >/dev/null && python3 -I -m json.tool .coldpass/cycles.json >/dev/null && echo ok; ls <guest-path> /inputs 2>&1 | head
cd <guest-path> && <env-var>=<guest-path> timeout 900 python3 -I <guest-path> --validate --replay 2>&1 | tail -c 4000
cd <guest-path> && <env-var>=<guest-path> timeout 900 python3 -I <guest-path> --validate --replay > <guest-path> 2>&1; echo rc=$?; python3 -I -c '
import json,re
t=open("<guest-path>").read()
print(len(t)); print(t[:600])
' ; grep -o '"valid": *[a-z]*\|"outcome": "[a-z_]*"\|"verdict": "[a-z]*"\|"error": "[^"]\{0,200\}' <guest-path> | head -20
cd <guest-path> && python3 -I -c '
import json
t=open("<guest-path>").read()
d=json.loads(t[t.index("{"):])
print(d.get("valid")); print(len(d.get("problems",[])))
for p in d.get("problems",[]): print("PROBLEM:", p[:150])
print(list(d.keys()))
'
grep -n "expected_returncode\|expected_exit_meaning\|def _example\|returncode" <guest-path> | head -30
sed -n 1520,1560p <guest-path> | cut -c1-260
cd <guest-path> && cat > .coldpass/report.json <<'EOF'
{"outcome":"working","fault_owner":"repo","operation":"Run the hound MCP server over stdio and call mcp_smart_fetch to fetch https://example.com/ (HTTP tier), returning page content","operation_scope":"Linux guest, Python 3.12.3 venv at <guest-path>, hound-mcp 12.4.1 installed from the pinned workspace source with [all] extras; HTTP-only fetch, no stealth browser used; network access required","steps":["pip install hound-mcp[all] (installed from the pinned workspace source instead of PyPI, to test the pinned code)","Browser download step for chromium (SKIPPED: task forbids downloading a browser; browser tier not exercised)","hound (stdio MCP server; initialize then tools/call mcp_smart_fetch url=https://example.com/ cache_ttl=0 returned 200 OK, fetcher_used=http, title Example Domain)","hound -v (version 12.4.1, reports 13.2.0 available)","python -m pytest tests/ -v --tb=short (CI command; 762 passed, 1 failed)"],"deviations":["pyproject.toml declares mcp>=1.27.0 with no upper bound; pip resolved mcp 2.3.0, and hound crashed at startup with AttributeError: 'Server' object has no attribute 'list_tools' (src/master_fetch/server.py:2987). Worked around in the venv by pinning mcp>=1.27.0,<2 (mcp 1.30.0); no product source edited.","Installed from the pinned workspace source (12.4.1) rather than PyPI (which provides 13.2.0), to test the pinned code.","Skipped the chromium browser download because the task forbids it; the CDP endpoint http://<internal-ip>:9226/json/version returned HTTP 200 (Chrome/155.0.8059.84) but no browser-dependent operation was run through it.","Tool names are exposed as mcp_smart_fetch, mcp_smart_crawl, mcp_screenshot, mcp_smart_search (README says smart_fetch etc.); used the server's actual names.","Test suite: tests/test_envelope.py::TestFreshness::test_no_fetched_at_falls_back_to_today fails on the VM clock (2026-10-28): it hardcodes a 2026-07-20 publish date and asserts age <= 10 days, but the actual age is 100 days. This is a date-dependent test defect in the repo, not a product fault.","Patchright printed a non-fatal Playwright-was-updated banner on stderr during fetch because no browser is installed."],"stuck_step":null,"main_blocker":null,"plain_failure":"Validation reports valid=false solely because the documented CI pytest example exits 1 on the date-dependent test in tests/test_envelope.py; the checker does not accept a nonzero example exit unless the source documents it, and the CI yml does not. The smart_fetch probe and README claim both replay successfully.","suggested_fix":"(1) Pin mcp<2 in pyproject.toml (or port server.py to the mcp 2.x API) so pip does not resolve an incompatible major version. (2) Make test_no_fetched_at_falls_back_to_today compute its expected age relative to the fixed reference date or freeze the clock, so it does not depend on the current date."}
EOF
python3 -I -m json.tool .coldpass/report.json >/dev/null && echo report_ok; ls -a; ls -d build *.egg-info .pytest_cache 2>/dev/null; echo done
cd <guest-path> && rm -rf .pytest_cache && <env-var>=<guest-path> timeout 900 python3 -I <guest-path> --validate --replay > <guest-path> 2>&1; echo rc=$?; python3 -I -c '
import json
t=open("<guest-path>").read()
d=json.loads(t[t.index("{"):])
print("valid=",d.get("valid")); [print("PROBLEM:",p[:120]) for p in d.get("problems",[])]'; rm -rf .pytest_cache; ls -a | head -3