MCP Observatory
2m ago Tracked 60,024 Last 2m ago 24h 2,459

Code analysis

static source read inferred

Static code-analysis findings — hidden prompt content in shipped skill files, committed secrets, dynamic-exec sinks, and suspicious call-home endpoints — across the analyzed catalogue. Heuristic, pure, no code executed; every row deep-links to its source. Click a kind to filter.

analysis coverage 86% of 59,698 analyzable servers
51224 analyzed
7479 re-analysis due
995 not analyzable
0 not yet analyzed
4849 source gone
not analyzable
795 too large 200 no source

Running analyzer v33. The scanner changelog explains what each version detects and when it changed.

committed secrets
  1. comisai/comis 62
  2. Vaibhav91one/assay 36
  3. Sertelegger/holdfast 35
  4. Cepeppe/handoff-mcp 31
  5. clidey/whodb 30
dynamic-exec sinks
  1. zilet/cairn 160
  2. abhishekpandaOfficial/XOne 149
  3. talirezun/the-curator 84
  4. TNLegend/PortaMCP 60
  5. mayowa-kalejaiye/Failures 55
suspicious endpoints
  1. nordsym/apiclaw 156
  2. tourhabk-ui/pos 115
  3. CarlxsMG/mcpbridge 90
  4. ergoveritas1-alt/certscore.ai 78
  5. JSONbored/loopover 68
credentials in logs
  1. gosswerner1-max/lazyrelay 8
  2. Hekkova/hekkova-mcp 6
  3. pendpost/pendpost 6
  4. Canonry/canonry 5
  5. KevinRabun/judges 5
over-broad OAuth scopes
  1. khill1269/servalsheets 82
  2. oneKn8/smart-mcps 26
  3. pipeshub-ai/pipeshub-ai 26
  4. lolxl/safespace-mcp 24
  5. BrianArfi/ai-second-brain 23
suspicious bundled scripts
  1. Dicklesworthstone/mcp_agent_mail 11
  2. syed-maaz/work-intelligence-mcp 7
  3. frankieramirez/ripen 6
  4. horustechltd/horus-flow-mcp 4
  5. itqanlab/agent-toolkit 4
ships skill files
  1. Amey-Thakur/AI-SKILLS 200
  2. ArasaniRohithReddy/LearningOS 200
  3. MANVENDRA-github/agentry 200
  4. Sandeeprdy1729/timps-swarm 200
  5. agentic-flow 200
code findings 15 shown
  1. HIGH dynamic exec openlegion-ai/openlegion __import__() backend.client.containers.get.side_effect = __import__(
    openlegion-ai-openlegion-24efd6e/tests/ test_runtime.py :134
  2. MEDIUM over-broad oauth scope openlegion-ai/openlegion https://mail.google.com/ for host in ("https://mail.google.com/inbox",
    openlegion-ai-openlegion-24efd6e/tests/ test_referer_picker.py :79
  3. HIGH dynamic exec openlegion-ai/openlegion __import__() router = __import__(
    openlegion-ai-openlegion-24efd6e/tests/ test_oauth_integrations_endpoints.py :77
  4. HIGH committed secret openlegion-ai/openlegion OpenAI key sk-sec…(31 chars, redacted)
    openlegion-ai-openlegion-24efd6e/tests/ test_oauth_integrations.py :496
  5. MEDIUM over-broad oauth scope openlegion-ai/openlegion https://mail.google.com/ await mgr.navigate("a1", "https://mail.google.com/")
    openlegion-ai-openlegion-24efd6e/tests/ test_navigate_referer.py :62
  6. HIGH committed secret openlegion-ai/openlegion OpenAI key sk-sup…(27 chars, redacted)
    openlegion-ai-openlegion-24efd6e/tests/ test_mesh.py :1104
  7. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.99 302, headers={"location": "https://93.184.216.99/elsewhere"},
    openlegion-ai-openlegion-24efd6e/tests/ test_mcp_oauth.py :179
  8. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.50 AS_ISSUER = "https://93.184.216.50"
    openlegion-ai-openlegion-24efd6e/tests/ test_mcp_oauth.py :26
  9. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.34 MCP_URL = "https://93.184.216.34/mcp"
    openlegion-ai-openlegion-24efd6e/tests/ test_mcp_oauth.py :25
  10. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.99 url="https://93.184.216.99/mcp", agents=["*"],
    openlegion-ai-openlegion-24efd6e/tests/ test_mcp_gateway.py :317
  11. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.34 url="https://93.184.216.34/mcp",
    openlegion-ai-openlegion-24efd6e/tests/ test_mcp_gateway.py :133
  12. HIGH committed secret openlegion-ai/openlegion Anthropic key sk-ant…(48 chars, redacted)
    openlegion-ai-openlegion-24efd6e/tests/ test_lifecycle.py :98
  13. HIGH committed secret openlegion-ai/openlegion Anthropic key sk-ant…(48 chars, redacted)
    openlegion-ai-openlegion-24efd6e/tests/ test_intent.py :93
  14. HIGH committed secret openlegion-ai/openlegion Anthropic key sk-ant…(54 chars, redacted)
    openlegion-ai-openlegion-24efd6e/tests/ test_deep_redact_blind_spots.py :30
  15. MEDIUM suspicious endpoint openlegion-ai/openlegion 93.184.216.34 url="https://93.184.216.34/mcp",
    openlegion-ai-openlegion-24efd6e/tests/ test_connectors_remote.py :41