ClawHub
Back to skill

Security audit

travel-search-ru

Security checks for vulnerabilities and agentic risk

Overview

The skill coherently provides travel searches through a disclosed external service, without booking purchases or accessing local credentials.

Searches share destinations, dates, traveler counts or ages, budgets, and preferences with mcp.botclaw.ru. Server retention is unspecified, so omit names, passport details, payment information, and credentials. Verify booking-link destinations, final prices, and availability before purchasing; some flight and train results are cached.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (14)

Env Variable Harvesting

High
Category
Data Exfiltration
Confidence
60% confidence
Finding

Code enumerates, copies, or searches environment variables for secrets. Bulk environment access can collect credentials unrelated to the skill's stated purpose.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 828)May include surrounding context.

python
"t.McpClient = C\n"
            "sys.exit(t.main(['search-tours', '--input', '{{}}']))\n"
        ).format(scripts=str(SCRIPTS), endpoint=self.server.endpoint)
        env = dict(os.environ, PYTHONIOENCODING="cp1252")
        proc = subprocess.run(
            [sys.executable, "-c", script], capture_output=True, env=env, timeout=30
        )

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: env, file_read, mcp, network, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 58)May include surrounding context.

md
If children are in the party and ages are unknown, **ask for ages** before presenting bookable family prices. Do not invent ages for pricing.

**Budget:** never auto-show offers that break a hard budget. Alternatives outside any hard constraint may appear only after **explicit user consent**, and only in a **separate labeled section**. Alternatives that satisfy every hard constraint (e.g. other hotels when the named ones do not fit) need no consent.

## When to use which command

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 829)May include surrounding context.

python
"sys.exit(t.main(['search-tours', '--input', '{{}}']))\n"
        ).format(scripts=str(SCRIPTS), endpoint=self.server.endpoint)
        env = dict(os.environ, PYTHONIOENCODING="cp1252")
        proc = subprocess.run(
            [sys.executable, "-c", script], capture_output=True, env=env, timeout=30
        )
        self.assertEqual(proc.returncode, 0, msg=proc.stderr)

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 914)May include surrounding context.

python
def test_timeout_exits_1(self):
        _STATE.delay_seconds = 2.0
        client = travel_search.McpClient(endpoint=self.server.endpoint, timeout=0.2)
        with self.assertRaises(travel_search.McpError):
            client.list_tools()
        # CLI path

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 929)May include surrounding context.

python
)
        # simpler: patch DEFAULT and construct via setUp override
        class SlowClient(self._orig_client):
            def __init__(self, endpoint=self.server.endpoint, timeout=0.2):
                super(SlowClient, self).__init__(endpoint=endpoint, timeout=timeout)

        travel_search.McpClient = SlowClient

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 922)May include surrounding context.

python
"C",
            (travel_search.McpClient,),
            {
                "__init__": lambda self, endpoint=self.server.endpoint, timeout=0.2: (
                    travel_search.McpClient.__init__(self, endpoint=endpoint, timeout=timeout)
                )
            },

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 1830)May include surrounding context.

python
cache_dir.is_dir() or True,
            "local caches may or may not exist; exclusion is still required",
        )
        proc = subprocess.run(
            ["npm", "pack", "--dry-run", "--json"],
            cwd=str(ROOT),
            capture_output=True,

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 1884)May include surrounding context.

python
_STATE.stream_mode = "sse_keepalive_only"
        _STATE.stream_hold_seconds = 5.0
        _STATE.stream_keepalive_interval = 0.05
        timeout = 0.6
        client = travel_search.McpClient(
            endpoint=self.server.endpoint, timeout=timeout
        )

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 1910)May include surrounding context.

python
_STATE.stream_hold_seconds = 3.0
        # Trickle faster than socket inactivity so readline stays blocked.
        _STATE.stream_keepalive_interval = 0.02
        timeout = 0.20
        client = travel_search.McpClient(
            endpoint=self.server.endpoint, timeout=timeout
        )

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 1934)May include surrounding context.

python
_STATE.stream_mode = "json_slow_trickle"
        _STATE.stream_hold_seconds = 3.0
        _STATE.stream_keepalive_interval = 0.02
        timeout = 0.20
        client = travel_search.McpClient(
            endpoint=self.server.endpoint, timeout=timeout
        )

Unbounded Resource Access

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill allows unbounded resource consumption (API calls, storage, compute). Without rate limits or quotas, a compromised or misbehaving agent can cause denial-of-service or cost overruns.

Content

Scanner excerpt · tests/test_travel_search.py (reported line 1958)May include surrounding context.

python
_STATE.stream_mode = "oversized_declared_trickle"
        _STATE.stream_hold_seconds = 3.0
        _STATE.stream_keepalive_interval = 0.02
        timeout = 0.20
        client = travel_search.McpClient(
            endpoint=self.server.endpoint, timeout=timeout
        )