Back to skill

Security audit

Venue Polling

Security checks for vulnerabilities and agentic risk

Overview

The skill is mostly coherent for venue polling and booking debugging, but it ships a live-looking authentication token and can automatically submit a booking order without confirmation.

Review before installing. Do not run the bundled polling script unless you intend it to place bookings automatically, and remove or rotate the exposed token-user credential first. The safer posture would be to load credentials from the user's environment, disable auto-booking by default, require explicit confirmation before createOrder, and limit invocation to explicit venue-booking tasks.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/venue_polling.py:20
Finding

Hard-Coded Reusable Authentication Token

Content
View full analysis
Remediation
View remediation
`. 5. Add secret-scanning checks to pre-commit hooks and continuous integration. 6. Ensure tokens are short-lived, narrowly scoped, revocable, and bound to the minimum API permissions required. 7. Review remote-service logs for use of the exposed token and invalidate related sessions if suspicious activity is found. ]]>

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/venue_polling.py:21
Finding

Automatic Order Creation Enabled Without Interactive Confirmation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (10)

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The script is described as a polling/helper tool, but AUTO_BOOK is enabled by default and the code includes a live order-submission path. In an agent skill context, this mismatch is dangerous because a user or higher-level agent may invoke it expecting read-only behavior, yet it can perform an external transactional action using the embedded token and signing flow.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

This code path constructs order data, signs the request, and sends a real booking POST to the production booking endpoint. That exceeds a passive polling/helper scope and creates a risk of unauthorized or unintended reservations, especially when run by an automated agent with access to valid credentials.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
91% confidence
Finding

The skill advertises operational guidance for reading local files and interacting with live booking APIs, but it does not declare any explicit tool scope such as permissions or allowed-tools. That omission weakens least-privilege controls because an agent may be permitted to use file and network capabilities implicitly, making it harder to constrain or audit potentially sensitive actions like reading key material or sending signed requests.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The default prompt says to use the skill to 'inspect, run, or modify' the venue polling and booking-signature workflow, which is broad enough to activate on generic debugging or maintenance requests rather than narrowly scoped booking tasks. In a skill that touches signed booking flows and replay/verification helpers, over-broad activation increases the chance the agent is invoked in unintended contexts and exposes sensitive operational logic or enables risky actions.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The reference explicitly documents that an embedded private key appears in config/signKey.js, which is sensitive credential material and should not be present in a client-distributed bundle. In the context of a skill focused on inspecting order-signing and replay helpers for booking APIs, surfacing this without a clear warning or guidance increases the chance of misuse, unauthorized request signing, and abuse of the booking flow.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill description says this script is for bundled signature replay and public-key verification helpers, but this helper derives the public key by opening rsa_private_key.pem and loading the private key material. Requiring private-key access is not justified for a public-key verification utility and expands the capability beyond the stated purpose.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

AUTO_BOOK defaults to True, so the first detected available venue can immediately trigger a booking request without any final confirmation. In an automation setting this can cause unintended purchases or reservations, and the presence of a hard-coded token makes the action easier to trigger if the script is reused or invoked indirectly.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
80% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/venue_polling.py (reported line 119)May include surrounding context.

python
sign_headers = build_sign_headers(order_data)
        headers.update(sign_headers)

        response = requests.post(API_ORDER_URL, headers=headers, json=order_data, timeout=15)
        result = response.json()

        if result.get("code") == 200 and result.get("success"):

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
76% confidence
Finding

The request body contains Chinese-language values such as the area name and sport name embedded directly in the sample payload. In isolation this suggests a locale-specific constraint with no accompanying natural-language indication that users can choose another language or that the skill is intentionally region-specific.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
88% confidence
Finding

The User-Agent includes 'Language/zh_CN', which forces a specific locale in a natural-language/config string. There is no indication that the locale is optional, user-configurable, or required for a justified region-specific purpose.

Content

No source excerpt is available for this finding.