{ "run_id": "voodoomods--hypgraph-claude-haiku-5-5-1793158439260896154", "date": "2026-10-28T03:37:07.770714+00:00", "repo": "voodoomods/hypgraph", "kind": "mcp", "source": "https://github.com/voodoomods/hypgraph", "commit": "cb6b9745076a2ce216c1a7bbcdfdf60554ad4bcb", "model": "claude-haiku-5-5", "model_version": "claude-haiku-5-5", "test_version": "readme-c03644542c69939a9374958136a8b3431f7456269d3fb4b3c433785f59cfd637", "pass": true, "outcome": "pass", "time_to_working_s": 157.76918365500023, "total_s": 188.50979852676392, "steps": [ "cd && pwd && ls -la && git status 2>&1 | head -3; id -u; python3 --version; cat README.md 2>/dev/null | head -150", "echo $CDP_URL; curl -s -m 10 -i $CDP_URL/json/version; echo; ls ; ls /inputs 2>&1", "cd && sed -n 150,400p README.md | cut -c1-400", "cd && grep -n -i \"install\\|quick\\|pip \\|python\" README.md | cut -c1-300 | head -40; grep -n \"requires-python\" pyproject.toml; ls .coldpass dev src tests", "mkdir -p && cd && python3 -m venv && install -q --upgrade pip 2>&1 | tail -3; install \"\" 2>&1 | tail -5; --help 2>&1 | head -60", "mkdir -p && cd && demo 2>&1 | head -40; echo \"rc=${PIPESTATUS[0]}\"; ls -la ", "cd && H=; $H --db research.hypgraph report 2>&1 | head -30; echo ---; $H bound 0 50 --exposure 1e6 2>&1; echo \"rc=$?\"; echo ---; $H --db research.hypgraph refusals 2>&1 | head -15", "cd && grep -n \"hypgraph bound\\|hypgraph demo \\|no dependencies\\|pip install hypgraph\" README.md | cat -A | cut -c1-250; echo ---; grep -rn \"def bound\\|def _bound\\|\\\"bound\\\"\" src/hypgraph/*.py | head; echo ---; grep -rln -i \"cron\\|schedule\\|webhook\\|while True\\|--once\\|--now\" src | head", "cd && grep -n -i \"cron\\|schedule\\|webhook\\|while True\\|--once\\|--now\" src/hypgraph/server.py src/hypgraph/coord.py src/hypgraph/bundle.py | cut -c1-220 | head -20; echo ---; grep -n \"^def \\|^async def \" src/hypgraph/mcp.py | head -5; sed -n 440,460p src/hypgraph/mcp.py | cut -c1-200", "cd && sed -n 1,12p src/hypgraph/coord.py | cut -c1-200; sed -n 232,250p src/hypgraph/coord.py | cut -c1-200; grep -n \"coord\" src/hypgraph/cli.py | head -5", "cd && grep -n \"hypgraph report \\|hypgraph refusals\" README.md | cat -A | cut -c1-200; cat > .coldpass/probe.json <<'EOF'\n{\"kind\":\"mcp\",\"argv\":[\"\",\"--db\",\"\",\"--as\",\"claude\",\"mcp\"],\"cwd\":\"\",\"tool\":\"bound\",\"arguments\":{\"k\":0,\"n\":50,\"exposure\":1000000},\"expected_text\":\"0 of 50\"}\nEOF\npython3 -I -c 'import json;print(json.load(open(\"\"))[\"tool\"])'", "cd && cat > .coldpass/claims.json <<'EOF'\n[{\"kind\":\"claim\",\"claim\":\"bound reports what 0 of 50 supports for a never claim\",\"source\":\"`hypgraph bound 0 50 --exposure 1e6` (or the MCP tool `bound`) answers\",\"input\":\"k=0 n=50 exposure=1e6\",\"expected\":\"rate 0% with 95% interval up to 5.8%\",\"probe_kind\":\"mcp\",\"argv\":[\"\",\"--db\",\"\",\"--as\",\"claude\",\"mcp\"],\"tool\":\"bound\",\"arguments\":{\"k\":0,\"n\":50,\"exposure\":1000000},\"expected_text\":\"0 of 50\",\"cwd\":\"\"},\n{\"kind\":\"claim\",\"claim\":\"hypgraph bound prints the 95% upper bound for 0 of 50\",\"source\":\"`hypgraph bound 0 50 --exposure 1e6` (or the MCP tool `bound`) answers\",\"input\":\"0 50 --exposure 1e6\",\"expected\":\"0 of 50 observed with interval up to 5.8%\",\"probe_kind\":\"cli\",\"argv\":[\"\",\"bound\",\"0\",\"50\",\"--exposure\",\"1e6\"],\"cwd\":\"\",\"expected_text\":\"0 of 50 observed: rate 0% (95% interval 0% to 5.8%)\"},\n{\"kind\":\"claim\",\"claim\":\"hypgraph report generates the facts from the record\",\"source\":\"hypgraph report # the facts, generated from the record\",\"input\":\"research.hypgraph from demo\",\"expected\":\"a facts section listing claims and results\",\"probe_kind\":\"cli\",\"argv\":[\"\",\"--db\",\"\",\"report\"],\"cwd\":\"\",\"expected_text\":\"Facts from the research graph\"}]\nEOF\ncat > .coldpass/examples.json <<'EOF'\n[{\"kind\":\"example\",\"claim\":\"demo builds a worked investigation, and refusals are logged\",\"source\":\"README.md Try it / Every refusal is also logged\",\"input\":\"research.hypgraph from hypgraph demo\",\"expected\":\"refusals listed per author\",\"argv\":[\"\",\"--db\",\"\",\"refusals\"],\"cwd\":\"\",\"expected_text\":\"Overclaims caught\"}]\nEOF\n= python3 -I --validate --replay 2>&1 | tail -60", "cd && cat > .coldpass/cycles.json <<'EOF'\n{\"detected\":false,\"reason\":\"Inspected README.md and src/hypgraph entry points (cli.py commands, mcp.py tools, server.py, coord.py). No cron, scheduler, --once/--now, or polling loop is documented. The only while-True loops are the SSE stream in server.py (GET /api/graph/stream), a helper walking supersedes chains in coord.py, and a coord.py loop over replaced orders; none is a product trigger cycle.\",\"checks\":[]}\nEOF\ncat > .coldpass/report.json <<'EOF'\n{\"outcome\":\"working\",\"fault_owner\":null,\"operation\":\"Install HypGraph from the pinned source and run its README quickstart: hypgraph demo, report, bound, refusals, and the MCP bound tool\",\"operation_scope\":\"Linux guest, Python 3.12.3 venv at (README requires Python >=3.10). Read-only operations on the demo graph; no external services, browser, or credentials used. CDP endpoint responded HTTP 200 but was not needed.\",\"steps\":[\"python3 -m venv \",\"pip install \\\"hypgraph[mcp]\\\" (from pinned checkout )\",\"hypgraph demo (built research.hypgraph in , exit 0)\",\"hypgraph --db research.hypgraph report\",\"hypgraph bound 0 50 --exposure 1e6\",\"hypgraph --db research.hypgraph refusals\",\"hypgraph --db research.hypgraph --as claude mcp, MCP tool bound(k=0,n=50,exposure=1000000)\"],\"deviations\":[\"README says pip install hypgraph (PyPI); installed the pinned local checkout with the [mcp] extra instead so the tested code is the pinned commit.\",\"hypgraph demo writes to the current directory, so it was run in rather than the checkout to keep the workspace clean.\",\"hypgraph serve (the web UI) was not started; the README quickstart lists it but it is an interactive long-running server and was not needed to exercise the product.\"],\"stuck_step\":null,\"main_blocker\":null,\"plain_failure\":null,\"suggested_fix\":\"None required for this run. No README-documented cycle or scheduler exists, so no cycle check was run.\"}\nEOF\npython3 -I -c 'import json;[json.load(open(\"\"+f)) for f in [\"report.json\",\"cycles.json\",\"examples.json\",\"claims.json\",\"probe.json\"]];print(\"json ok\")'; = python3 -I --validate --replay 2>&1 | python3 -I -c 'import json,sys; d=json.load(sys.stdin); print(d[\"valid\"], d[\"problems\"], d[\"replay\"][\"status\"])' 2>&1 | tail -5" ], "stuck_step": null, "error_excerpt": null, "what_it_did": "Install HypGraph from the pinned source and run its README quickstart: hypgraph demo, report, bound, refusals, and the MCP bound tool", "suggested_fix": null, "tokens_in": 285273, "tokens_out": 10157, "cost_usd": 0.014579020000000002, "usage_complete": true, "checks": [ { "check": "operation", "status": "PASS", "details": { "working": true, "argv": [ "", "--db", "", "--as", "claude", "mcp" ], "kind": "mcp", "serverInfo": { "name": "hypgraph", "version": "0.7.1", "websiteUrl": "https://hypgraph.com" }, "tools": [ "overview", "find", "show", "start_investigation", "add", "link", "revise", "withdraw", "retract_link", "review", "next_experiments", "work_queue", "claim_work", "register_run", "report", "verify", "bound", "check_report", "refusal_log", "briefing", "issue_work", "my_work", "check_work", "return_work", "accept_work", "rework", "cancel_work", "completion", "stage", "settle_work", "cancel_moot", "schema" ], "tool_schema_bytes": 21264, "tool_result": { "_meta": { "fastmcp": { "wrap_result": true } }, "content": [ { "text": "0 of 50 observed: rate 0% (95% interval 0% to 5.8%)\n'never' here means: at most 5.8% \u2014 1 in 20 \u2014 at 95% confidence\n to bound it below 1 in 100: about 299 trials with none\n to bound it below 1 in 1,000: about 2,995 trials with none\n to bound it below 1 in 10,000: about 29,956 trials with none\nacross 1,000,000 uses: up to 58,155 not ruled out (95%)", "type": "text" } ], "isError": false, "structuredContent": { "result": "0 of 50 observed: rate 0% (95% interval 0% to 5.8%)\n'never' here means: at most 5.8% \u2014 1 in 20 \u2014 at 95% confidence\n to bound it below 1 in 100: about 299 trials with none\n to bound it below 1 in 1,000: about 2,995 trials with none\n to bound it below 1 in 10,000: about 29,956 trials with none\nacross 1,000,000 uses: up to 58,155 not ruled out (95%)" } }, "tool_result_checked": true, "scope": "product operation" } } ], "safety": { "version": "z2x-1", "rules_version": "z2x-1", "summary": "No problems traced to this repository code in these checks.", "expected": "Reviewed as expected for a test run: attempted writes that failed, so nothing was written.", "findings": [], "context": [], "limitations": [ "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." ], "static": { "dependency_note": "OSV scanned supported lockfiles; unpinned dependencies and unsupported formats are outside this check." }, "evidence_url": "/safety/?run=voodoomods--hypgraph-claude-haiku-5-5-1793158439260896154", "evidence_file": "/data/safety/voodoomods--hypgraph-claude-haiku-5-5-1793158439260896154.json.gz" }, "agent_execution": "native_agent", "agent_runtime": "claude-code", "agent_runtime_version": "2.1.310 (Claude Code)", "tool_transport": "mcp-assigned-vm", "platform": "linux", "reasoning": "low", "service_tier": null, "resources": { "cpus": 2, "memory_mib": 4096, "disk_mib": 32768 }, "native_tool_calls": 13, "native_bundle": null, "architecture": null, "backend": "hyperv", "platform_details": { "backend": "hyperv", "os": "linux" }, "platform_conflicts": null, "sandbox_source": null, "source_pin": { "commit": "cb6b9745076a2ce216c1a7bbcdfdf60554ad4bcb", "url": "https://github.com/voodoomods/hypgraph" }, "purpose": "initial", "profile": { "model": "claude-haiku-5-5", "reasoning": "low" }, "fault": { "owner": null, "status": "resolved", "reason": null }, "fix_id": null, "verifies_run_id": null, "hosting": null, "operation_scope": "product operation", "tool_calls": 13, "guest_setup": null, "runtime": null, "runtime_version": null, "runtime_revision": null, "source_revision": null, "evidence_version": null, "failure_origin": null, "evidence_strength": "readme_checked", "works": "yes", "install_s": null, "first_use_s": null, "mcp": "local", "needs": [], "needs_which": null, "agent_interface": [ "mcp" ], "license": "Apache-2.0", "repo_url": "https://github.com/voodoomods/hypgraph", "cost_to_try": "free", "hardware": null, "what_failed": null, "local_or_hosted": "local", "cycle_checks": [ { "kind": "cycle", "claim": "Scheduled or looping main behavior", "source": "repository entry points", "input": "source inspection", "expected": "identify the product trigger", "verdict": "yes", "outcome": "Inspected README.md and src/hypgraph entry points (cli.py commands, mcp.py tools, server.py, coord.py). No cron, scheduler, --once/--now, or polling loop is documented. The only while-True loops are the SSE stream in server.py (GET /api/graph/stream), a helper walking supersedes chains in coord.py, and a coord.py loop over replaced orders; none is a product trigger cycle.", "evidence": "Inspected README.md and src/hypgraph entry points (cli.py commands, mcp.py tools, server.py, coord.py). No cron, scheduler, --once/--now, or polling loop is documented. The only while-True loops are the SSE stream in server.py (GET /api/graph/stream), a helper walking supersedes chains in coord.py, and a coord.py loop over replaced orders; none is a product trigger cycle.", "command": "" } ], "claims": [ { "kind": "claim", "claim": "bound reports what 0 of 50 supports for a never claim", "source": "`hypgraph bound 0 50 --exposure 1e6` (or the MCP tool `bound`) answers", "input": "{\"tool\": \"bound\", \"arguments\": {\"k\": 0, \"n\": 50, \"exposure\": 1000000}}", "expected": "rate 0% with 95% interval up to 5.8%", "command": " --db --as claude mcp tools/call {\"tool\": \"bound\", \"arguments\": {\"k\": 0, \"n\": 50, \"exposure\": 1000000}}", "verdict": "yes", "outcome": "Real input produced the expected result: 0 of 50", "evidence": "\u20260 of 50 observed: rate 0% (95% interval 0% to 5.8%)\\n'never' here means: at most 5.8% \u2014 1 in 20 \u2014 at 95% confidence\\n to bound it below 1 in 100: about 299 trials with none\\n to bound it below 1 in 1,000: about 2,995 trials with none\\n to bound it below 1 in 10,000: about 29,956 trials with none\\nacross 1,000,000 uses: up to 58,155 not ruled out (95%)\", \"type\": \"text\"}], \"isError\": false, \"structuredContent\": {\"result\": \"0 of 50 observed: rate 0% (95% interval 0% to 5.8%)\\n'never' here means: at most 5.8% \u2014 1 in 20 \u2014 at 95% confidence\\n to bound it below 1 in 100: about 299 trials with non", "source_checked": true, "source_file": "README.md" }, { "kind": "claim", "claim": "hypgraph bound prints the 95% upper bound for 0 of 50", "source": "`hypgraph bound 0 50 --exposure 1e6` (or the MCP tool `bound`) answers", "input": "0 50 --exposure 1e6", "expected": "0 of 50 observed with interval up to 5.8%", "command": " bound 0 50 --exposure 1e6", "verdict": "yes", "outcome": "Real input produced the expected result: 0 of 50 observed: rate 0% (95% interval 0% to 5.8%)", "evidence": "0 of 50 observed: rate 0% (95% interval 0% to 5.8%)\n'never' here means: at most 5.8% \u2014 1 in 20 \u2014 at 95% confidence\n to bound it below 1 in 100: about 299 trials with none\n to bound it below 1 in 1,000: about 2,995 trials with none\n to bound it below 1 in 10,000: about 29,956 trials with none\nacross 1,000,000 uses: up to 58,155 not ruled out (95%)\n", "returncode": 0, "source_checked": true, "source_file": "README.md" }, { "kind": "claim", "claim": "hypgraph report generates the facts from the record", "source": "hypgraph report # the facts, generated from the record", "input": "research.hypgraph from demo", "expected": "a facts section listing claims and results", "command": " --db report", "verdict": "yes", "outcome": "Real input produced the expected result: Facts from the research graph", "evidence": "\u2026Facts from the research graph \u2014 INV1 How does cholera spread? Broad Street, 1854 (active)\n_Generated from the record at event 59; do not edit \u2014 regenerate it from the graph. Interpretation goes below this section._\n\n### Claims\n\n| claim | standing | review | verdicts | statement |\n|---|---|---|---|---|\n| CL1 v2 | active | checked | whitehead: objection (RV1, on v1); whitehead: resolved (RV2, on v1); whitehead: replicated (RV3, on v1); whitehead: checked (RV4, on v2) | The Golden Square outbreak was spread by water from the Broad Street pump: those who drank it died, while 5 of the 535 workhouse", "returncode": 0, "source_checked": true, "source_file": "README.md" } ], "claims_note": null, "example_checks": [ { "kind": "example", "claim": "demo builds a worked investigation, and refusals are logged", "source": "README.md Try it / Every refusal is also logged", "input": "research.hypgraph from hypgraph demo", "command": " --db refusals", "evidence": "EXCLUSION_WITHOUT_POWER \u00d71, EVIDENCE_MISSING \u00d71, SCOPE_MISSING \u00d71 |\n\n| rule | times |\n|---|---|\n| `EXCLUSION_WITHOUT_POWER` | 1 |\n| `EVIDENCE_MISSING` | 1 |\n| `SCOPE_MISSING` | 1 |\n| `ABSOLUTE_WITHOUT_COUNT` | 1 |\n\n### Latest\n\n- 1854-11-20T10:00 **you**: \u201cThe water-company comparison rules out the miasma theory.\u201d \u2014 EXCLUSION_WITHOUT_POWER \u2192 not fixed\n- 1854-09-08T10:00 **you**: \u201cRemoving the pump handle ended the epidemic.\u201d \u2014 EVIDENCE_MISSING, SCOPE_MISSING \u2192 not fixed\n- 1854-09-07T10:00 **you**: \u201cEvery death in the outbreak was near the Broad Street pump.\u201d \u2014 ABSOLUTE_WITHOUT_COUNT \u2192 not fixed", "deferred": false, "expected": "refusals listed per author", "returncode": 0, "verdict": "yes", "outcome": "command completed successfully" } ], "does_what_it_says": "yes", "maintenance": { "last_commit_at": "2026-10-28T01:52:53Z", "commits_90d": 151, "merged_prs_90d": 28, "open_issues": 0, "open_prs": 0, "reply_share": null, "reply_threads": 0, "median_reply_hours": null, "last_release_at": null, "archived": false, "verdict": "maintained", "note": "actively maintained: last commit 0 days ago; 151 commits and 28 merged PRs in 90 days; 0 open issues, 0 open PRs; no releases" }, "maintenance_verdict": "maintained", "maintenance_note": "actively maintained: last commit 0 days ago; 151 commits and 28 merged PRs in 90 days; 0 open issues, 0 open PRs; no releases", "maintained": "active", "maintained_note": "last commit 0 days ago", "last_commit_at": "2026-10-28T01:52:53Z", "last_release_at": null, "open_issues": null, "answered_issues": null, "platforms_claimed": [ "windows" ], "topic": "mcp-server", "agent_interfaces": [ { "interface": "mcp", "evidence": "independent checker", "observed": true, "operation_checked": true } ], "agent_interfaces_note": "README metadata is a discovery hint, not pinned documentation proof. Observed interfaces were exercised in this run. An empty list means unknown.", "mcp_context": { "tool_count": 32, "schema_bytes": 21264, "tokens_approx": 5316, "estimate_method": "ceil(compact UTF-8 tools/list JSON bytes / 4)", "scope": "returned tools/list page only; excludes client framing and tool results" }, "tested_at": "2026-10-28T03:37:07.770714+00:00", "tested_commit": "cb6b9745076a2ce216c1a7bbcdfdf60554ad4bcb", "tested_with": { "agent": "claude-code", "agent_version": "2.1.310 (Claude Code)", "model": "claude-haiku-5-5", "model_route": "claude-haiku-5-5", "reasoning": "low", "os": "linux", "backend": "hyperv", "transport": "mcp-assigned-vm" }, "setup_identity": { "repo": "voodoomods/hypgraph", "source": "https://github.com/voodoomods/hypgraph", "commit": "cb6b9745076a2ce216c1a7bbcdfdf60554ad4bcb", "source_revision": null, "os": "linux", "architecture": null, "backend": "hyperv", "model": "claude-haiku-5-5", "model_version": "claude-haiku-5-5", "agent_execution": "native_agent", "agent_runtime": "claude-code", "agent_runtime_version": "2.1.310 (Claude Code)", "tool_transport": "mcp-assigned-vm", "runtime": null, "runtime_version": null, "runtime_revision": null, "sandbox_source": null, "guest_setup": null, "source_pin": { "commit": "cb6b9745076a2ce216c1a7bbcdfdf60554ad4bcb", "url": "https://github.com/voodoomods/hypgraph" }, "platform_details": { "os": "linux", "backend": "hyperv" }, "platform_conflicts": null, "reasoning": "low", "service_tier": null, "resources": { "cpus": 2, "memory_mib": 4096, "disk_mib": 32768 }, "operation": "Install HypGraph from the pinned source and run its README quickstart: hypgraph demo, report, bound, refusals, and the MCP bound tool", "operation_scope": "product operation", "hosting": null, "kind": "mcp", "test_version": "readme-c03644542c69939a9374958136a8b3431f7456269d3fb4b3c433785f59cfd637", "evidence_version": null }, "other_model": null, "other_run_id": null, "models_agree": null, "furthest_model": null, "transcript_url": "/transcripts/voodoomods--hypgraph-claude-haiku-5-5-1793158439260896154.txt", "report_url": "/reports/voodoomods--hypgraph-claude-haiku-5-5-1793158439260896154.json" }