Dos Kernel — the complete, unedited output of the deterministic mcptrustchecker engine v1.9.0, scanned . Every finding, capability tag and score component below is exactly what the engine produced — no AI, no post-processing.
{
"tool": {
"name": "mcptrustchecker",
"version": "1.9.0",
"methodologyVersion": "mcptrustchecker-1.9"
},
"target": {
"id": "dos-kernel",
"source": {
"kind": "package",
"origin": "dos-kernel"
},
"server": {
"name": "dos-kernel"
}
},
"grade": "A",
"score": {
"score": 93,
"threatScore": 100,
"grade": "A",
"band": "A",
"categorySubtotals": {
"injection": 0,
"exfiltration": 0,
"permissions": 0,
"supply-chain": 0,
"network": 0,
"hygiene": 0
},
"vector": [
{
"kind": "client",
"term": "capability-exposure",
"level": "high",
"label": "capability blast radius (high) — client exposure if the model is manipulated",
"appliedPenalty": 6
},
{
"kind": "client",
"term": "verification-discount",
"level": "repo",
"label": "publisher verification (public source) — no provenance, but the source is public and inspectable",
"appliedPenalty": 1
},
{
"kind": "client",
"term": "coverage-honesty",
"level": "source",
"label": "inspection depth (source) — how much of the target the scan could see",
"appliedPenalty": 0
}
],
"gatesFired": [],
"methodologyVersion": "mcptrustchecker-1.9"
},
"capability": {
"level": "high",
"reasons": [
"ingests untrusted external content (a prompt-injection entry point)",
"reads sensitive or local data",
"has a read → egress path (a data-exfiltration surface)",
"can execute shell commands or code"
],
"tags": [
"untrusted-input",
"sensitive-source",
"code-exec"
]
},
"coverage": {
"level": "source",
"inputs": {
"toolSurface": true,
"implementationSource": true,
"packageMetadata": true,
"liveTransport": false
},
"caveats": [
"Tools were statically extracted from the published source (11 recovered), not enumerated from a running server. Tool-poisoning, Unicode-smuggling, capability and toxic-flow analysis ran on this inferred surface, but a mis-parsed registration could be missed or mis-attributed, so tool-derived findings are capped below “confirmed”. To grade the real runtime surface, scan the running server: --command \"npx -y <package>\"."
]
},
"findings": [
{
"ruleId": "MTC-SRC-006",
"title": "Reads a sensitive credential path or dumps the environment (src/dos/call_shape.py)",
"category": "exfiltration",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/call_shape.py:280`): Reading private keys / cloud credentials, or serializing the whole environment, is a sensitive-data source that becomes exfiltration when combined with any egress. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/call_shape.py"
},
"evidence": "collides with `.ssh/id_rsa`. Returns the offending glob (un-normalized) so the refusal can name it; None = the t",
"owasp": "LLM02:2025 Sensitive Information Disclosure",
"data": {
"rule": "MTC-SRC-006",
"file": "src/dos/call_shape.py",
"line": 280,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/__init__.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/__init__.py:24`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/__init__.py"
},
"evidence": "rnel — ACR Plane ①) spawn()/reap() -> dos.run_id (the correlation spine across subprocess boundaries)",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/__init__.py",
"line": 24,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/cli.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/cli.py:769`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/cli.py"
},
"evidence": "e run\") start_sha = subprocess.run( [\"git\", \"-C\", str(work), \"rev-parse\", \"--short\", \"HEAD\"], check=",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/cli.py",
"line": 769,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/decisions_tui.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/decisions_tui.py:51`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/decisions_tui.py"
},
"evidence": "try: p = subprocess.run(cmd, input=text.encode(\"utf-8\"), timeout=5) if p.returncode == 0:",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/decisions_tui.py",
"line": 51,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/_effect_gate.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/_effect_gate.py:358`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/_effect_gate.py"
},
"evidence": "try: proc = subprocess.run( [\"git\", *args], cwd=self._root(),",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/_effect_gate.py",
"line": 358,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/ci_status.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/ci_status.py:405`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/ci_status.py"
},
"evidence": "\"\" try: p = subprocess.run( [\"gh\", *args], capture_output=True, text=Tru",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/ci_status.py",
"line": 405,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/content_diff.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/content_diff.py:394`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/content_diff.py"
},
"evidence": "try: p = subprocess.run( [\"git\", *args], capture_output=True, text=True,",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/content_diff.py",
"line": 394,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/llm_judge.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/llm_judge.py:103`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/llm_judge.py"
},
"evidence": "…` string. p = subprocess.run( cmd, shell=True, input=prompt.encode(\"utf-8\"",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/llm_judge.py",
"line": 103,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/os_acceptance.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/os_acceptance.py:130`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/os_acceptance.py"
},
"evidence": "try: p = subprocess.run( argv, capture_output=True, text=",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/os_acceptance.py",
"line": 130,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/provider_ledger.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/provider_ledger.py:452`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/provider_ledger.py"
},
"evidence": "try: p = subprocess.run( [\"git\", *args], capture_output=True, text=True,",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/provider_ledger.py",
"line": 452,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-002",
"title": "Shell/command execution in server code (src/dos/drivers/similarity_judge.py)",
"category": "permissions",
"severity": "high",
"confidence": "strong",
"description": "In the server's implementation (`src/dos/drivers/similarity_judge.py:143`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "src/dos/drivers/similarity_judge.py"
},
"evidence": "de(\"utf-8\") p = subprocess.run( cmd, shell=True, input=payload, capture_output=True, timeout=120,",
"owasp": "LLM05:2025 Improper Output Handling",
"data": {
"rule": "MTC-SRC-002",
"file": "src/dos/drivers/similarity_judge.py",
"line": 143,
"nonRuntime": false
}
},
{
"ruleId": "MTC-SRC-006",
"title": "Reads a sensitive credential path or dumps the environment in packaging/dev tooling (tests/test_call_shape.py)",
"category": "exfiltration",
"severity": "low",
"confidence": "heuristic",
"description": "In a packaging/dev/install script (shipped, but not the server runtime) (`tests/test_call_shape.py:95`): Reading private keys / cloud credentials, or serializing the whole environment, is a sensitive-data source that becomes exfiltration when combined with any egress. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.",
"remediation": "Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.",
"location": {
"kind": "server",
"name": "tests/test_call_shape.py"
},
"evidence": "ne=\"Write\", tree=(\".ssh/id_rsa\",)), {}, None) assert not v.admitted # --------------------------------------------",
"owasp": "LLM02:2025 Sensitive Information Disclosure",
"data": {
"rule": "MTC-SRC-006",
"file": "tests/test_call_shape.py",
"line": 95,
"nonRuntime": true
}
}
],
"toxicFlows": [],
"capabilities": [
{
"tool": "dos_verify",
"tags": [],
"reasons": {}
},
{
"tool": "dos_commit_audit",
"tags": [],
"reasons": {}
},
{
"tool": "dos_review",
"tags": [],
"reasons": {}
},
{
"tool": "dos_arbitrate",
"tags": [],
"reasons": {}
},
{
"tool": "dos_refuse_reasons",
"tags": [
"untrusted-input"
],
"reasons": {
"untrusted-input": [
"keyword \"browse\""
]
}
},
{
"tool": "dos_check_reason",
"tags": [],
"reasons": {}
},
{
"tool": "dos_recall",
"tags": [],
"reasons": {}
},
{
"tool": "dos_doctor",
"tags": [],
"reasons": {}
},
{
"tool": "dos_status",
"tags": [],
"reasons": {}
},
{
"tool": "dos_citation_resolve",
"tags": [],
"reasons": {}
},
{
"tool": "dos_answer",
"tags": [
"untrusted-input"
],
"reasons": {
"untrusted-input": [
"keyword \"fetch\""
]
}
}
],
"surfaceDigest": "63aa84c1890fcae243e848cc65d35c9eddcf7304906ff3650307bf3c956cd79e",
"stats": {
"tools": 11,
"prompts": 0,
"resources": 0,
"findingsBySeverity": {
"critical": 0,
"high": 11,
"medium": 0,
"low": 1,
"info": 0
}
}
}