github Python re-analysis due deep scan · partial (287/3,388 files)

Abilityai/trinity

Self-hosted AI Agents Platform supporting Claude Code, Codex, Gemini agents. Apache 2.0.

maintainer: Abilityai
licence: Apache-2.0
first seen: 2026-09-04
last seen: 2026-09-17
releases · 30d: 4
short id

risk 44/100 · heuristic grade

C elevated inferred analyzer v33 (current)

capability exposure inferred + 35
tool safety inferred + 12
trust mitigators mixed − 3

inferred mixed

The A–E grade is our heuristic synthesis — a "review this" prompt, not a verdict. Each factor is tagged by what backs it: attested (a verifiable record), reported (a third party's claim), or inferred (our own heuristic, e.g. permissions). See methodology.

grade last moved 1w ago · see ecosystem CVEs →

capability exposure grade factor +35

Inferred surface — each links to servers holding it:

vulnerabilities 0 CVEs

No known CVEs for this server.

tool safety 1 findings · grade factor +12

high dangerous code

committed secret: Slack token, Google API key, Anthropic key, GitHub fine-grained PAT, GitHub token, OpenAI key, Slack app token · dynamic exec: eval()/exec(), __import__ sink, __…

skills & danger signals github-tarball

prompt-surface shipped agent-instruction files + hidden-content / dangerous-code findings — quoted from the analyzed source

analyzed commit e38c1c0 · analyzer v33 · 4h ago

skills & prompt files 69

danger signals56

dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/test_rate_limits.py :77 exec(func_source, ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_1068_task_timeout_demotion.py :51 exec(compile(module, filename=str(_CHAT_SRC), mode="exec"), ns) # noqa: S102
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_1810_1814_version_truth.py :53 exec(snippet, ns)
dynamic code execution __import__ sink Abilityai-trinity-e38c1c0/tests/unit/test_2247_calendar_rot_detector.py :66 "PATH": __import__("os").environ.get("PATH", ""),
dynamic code execution __import__() Abilityai-trinity-e38c1c0/tests/unit/test_2323_machine_identities.py :96 mod = __import__(module_name, fromlist=["router"])
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_2380_install_provenance.py :1180 exec(snippet, ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_2380_provision_single_source.py :322 exec(snippet, ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_2691_public_url_reachability.py :73 exec(_slice(src, "_ASK_HOSTS = "), ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_926_version_endpoint.py :47 exec(snippet, ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_guardrails.py :132 exec(snippet, ns)
dynamic code execution eval()/exec() Abilityai-trinity-e38c1c0/tests/unit/test_session_inflight_sentinel.py :321 exec(compile(ast.Module(body=[fn], type_ignores=[]), "<ast>", "exec"), ns)
suspicious endpoint api.telegram.org Abilityai-trinity-e38c1c0/src/backend/adapters/telegram_adapter.py :35 TELEGRAM_API_BASE = "https://api.telegram.org"
suspicious endpoint api.telegram.org Abilityai-trinity-e38c1c0/src/backend/adapters/transports/telegram_webhook.py :151 url = f"https://api.telegram.org/bot{bot_token}/setWebhook"
suspicious endpoint t.me Abilityai-trinity-e38c1c0/src/backend/routers/telegram.py :121 bot_link=f"https://t.me/{bot_username}" if bot_username else None,
suspicious endpoint api.telegram.org Abilityai-trinity-e38c1c0/src/backend/routers/telegram.py :188 resp = await client.get(f"https://api.telegram.org/bot{bot_token}/getMe")
suspicious endpoint api.telegram.org Abilityai-trinity-e38c1c0/src/backend/services/telegram_media.py :25 TELEGRAM_API_BASE = "https://api.telegram.org"
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/test_mcp_validator_endpoint.py :159 {"e": {"type": "http", "url": "https://169.254.169.254/"}},
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_1951_slack_inbound_media.py :166 "https://169.254.169.254/latest/meta-data/", # cloud metadata
suspicious endpoint 1.2.3.4 Abilityai-trinity-e38c1c0/tests/unit/test_2380_install_provenance.py :726 "https://1.2.3.4",
suspicious endpoint 1.2.3.4 Abilityai-trinity-e38c1c0/tests/unit/test_736_a2a_outbound_edges.py :711 client, "GET", "https://1.2.3.4/x",
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_736_a2a_outbound_transport.py :369 return httpx.Response(307, headers={"location": "http://169.254.169.254/"})
suspicious endpoint 93.184.216.34 Abilityai-trinity-e38c1c0/tests/unit/test_736_a2a_outbound_transport.py :662 assert pinned == "https://93.184.216.34/a2a"
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_ent14_catalog_failopen.py :283 "redirect": lambda r: httpx.Response(302, headers={"location": "http://169.254.169.254/"}),
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_ent14_registry_fetch.py :387 return httpx.Response(status_code, headers={"location": "http://169.254.169.254/"})
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_ent346_skills_source_injection.py :147 "http://169.254.169.254/latest/meta-data", # SSRF classic
suspicious endpoint 93.184.216.34 Abilityai-trinity-e38c1c0/tests/unit/test_ent399_ipv6_origin.py :113 ("https://93.184.216.34/a2a", "https://[::ffff:93.184.216.34]/a2a",
suspicious endpoint 93.184.216.034 Abilityai-trinity-e38c1c0/tests/unit/test_ent399_ipv6_origin.py :129 assert same_origin("https://93.184.216.034/a2a", "https://93.184.216.34/a2a") is False
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_mcp_validator.py :529 cfg = _wrap({"r": {"type": "http", "url": "https://169.254.169.254/latest/meta-data/"}})
suspicious endpoint 169.254.169.254 (cloud metadata) Abilityai-trinity-e38c1c0/tests/unit/test_ssrf_skills_library.py :100 validate_skills_library_url("https://169.254.169.254/latest/meta-data/")
over-broad OAuth scope https://www.googleapis.com/auth/drive Abilityai-trinity-e38c1c0/src/backend/routers/credentials.py :271 "google": "openid email profile https://www.googleapis.com/auth/drive",
committed secret Slack token Abilityai-trinity-e38c1c0/docs/testing/phases/PHASE_24_CREDENTIAL_MANAGEMENT.md :173 xoxb-s…(21 chars, redacted)
committed secret Google API key Abilityai-trinity-e38c1c0/src/frontend/tests/unit/reportSummary.spec.js :130 AIzaSy…(39 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/test_settings.py :542 sk-ant…(41 chars, redacted)
committed secret GitHub fine-grained PAT Abilityai-trinity-e38c1c0/tests/test_settings.py :826 github…(34 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/test_subscription_auto_switch.py :25 sk-ant…(36 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/test_subscription_usage.py :22 sk-ant…(39 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/test_subscriptions.py :32 sk-ant…(36 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_1816_system_agent_adoption.py :1328 sk-ant…(45 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_1853_error_telemetry_salvage.py :66 sk-ant…(39 chars, redacted)
committed secret GitHub fine-grained PAT Abilityai-trinity-e38c1c0/tests/unit/test_2007_mcp_template_render.py :396 github…(43 chars, redacted)
committed secret Slack token Abilityai-trinity-e38c1c0/tests/unit/test_267_credential_key_rotation.py :36 xoxb-s…(17 chars, redacted)
committed secret GitHub token Abilityai-trinity-e38c1c0/tests/unit/test_2744_skills_adoption_alert_idempotency.py :78 ghp_AA…(40 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_agent.py :57 sk-ant…(43 chars, redacted)
committed secret GitHub fine-grained PAT Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_agent.py :71 github…(57 chars, redacted)
committed secret GitHub token Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_agent.py :271 ghp_98…(44 chars, redacted)
committed secret OpenAI key Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_agent.py :282 sk-222…(33 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_backend.py :65 sk-ant…(43 chars, redacted)
committed secret GitHub fine-grained PAT Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_backend.py :79 github…(57 chars, redacted)
committed secret GitHub token Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_backend.py :279 ghp_98…(44 chars, redacted)
committed secret OpenAI key Abilityai-trinity-e38c1c0/tests/unit/test_credential_sanitizer_backend.py :290 sk-222…(33 chars, redacted)
committed secret GitHub token Abilityai-trinity-e38c1c0/tests/unit/test_ent125_resilient_system_deploy.py :335 ghp_01…(42 chars, redacted)
committed secret Slack app token Abilityai-trinity-e38c1c0/tests/unit/test_ent435_secret_settings.py :378 xapp-1…(19 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_ent457_portal_completion_report.py :274 sk-ant…(37 chars, redacted)
committed secret Anthropic key Abilityai-trinity-e38c1c0/tests/unit/test_error_classifier_dict_body.py :116 sk-ant…(76 chars, redacted)
committed secret Slack token Abilityai-trinity-e38c1c0/tests/unit/test_slack_token_encryption.py :120 xoxb-2…(30 chars, redacted)
committed secret Slack token Abilityai-trinity-e38c1c0/tests/unit/test_slack_workspaces_encryption.py :118 xoxb-a…(23 chars, redacted)

embed badge readme-ready

[![MCP Observatory risk grade](https://mcpobservatory.com/servers/github:Abilityai/trinity/badge.svg)](https://mcpobservatory.com/servers/github:Abilityai/trinity/security)

Heuristic, inferred signals — false positives (legitimately powerful tools, forks, language ports) are expected. Treat each as "review this", not a verdict. See the ecosystem-wide picture on the security hub, or the fleet security of Abilityai.

Abilityai/trinity

skills & prompt files 69

danger signals56

Navigate

Stream

Features

Display