Back to skill

Security audit

InterCEPT SDR

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent SDR platform control reference, but it documents default administrator login, session-cookie use, and high-impact control endpoints without enough safeguards.

Install only for an iNTERCEPT instance you own or are explicitly authorized to operate. Rotate any default admin credentials before use, prefer scoped credentials or tokens, keep remote access behind HTTPS, avoid sending cookies over untrusted networks, and require explicit confirmation before kill, stop/start, recording, WiFi monitor mode, message sending, or remote-agent actions.

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
SKILL.md:16
Finding

Hardcoded Default Administrator Credentials

Content
View full analysis
]*value="\K[^"]+'` 3. Use the session cookie for subsequent requests ``` ### Technical Analysis The Skill explicitly instructs the Agent to authenticate with the fixed credentials `admin:admin`. These credentials are included in publicly readable Skill documentation and therefore cannot be treated as secrets. The CSRF token requirement does not mitigate this weakness. An attacker who can reach the iNTERCEPT server can request the login page, extract a valid CSRF token using the documented procedure, and authenticate with the default credentials. The resulting session cookie provides access to authenticated API operations. This authentication method exceeds minimum necessary privilege because the Skill directs all authenticated operations through a shared administrator account rather than using task-specific, least-privileged credentials. The API includes sensitive and disruptive functionality such as modifying settings, enabling wireless monitor mode, controlling remote agents, retrieving signal intelligence, and terminating running processes. ### Attack Path 1. The attacker discovers an accessible iNTERCEPT deployment, including one exposed through a local network, port forwarding, container mapping, or a remote `INTERCEPT_BASE_URL`. 2. The attacker requests the unauthenticated `/login` page. 3. The attacker extracts the CSRF token from the returned HTML. 4. The attacker submits `username=admin`, `password=admin`, and the extracted CSRF token to `POST /login`. 5. If the deployment retains the documented defaults, the server returns an authenticated administrator session cookie. 6. ...[truncated 788 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:13
Finding

Credentials and Session Cookies May Be Transmitted Over Plaintext HTTP

Content
View full analysis
]*value="\K[^"]+'` 3. Use the session cookie for subsequent requests ``` ```text ## Auth Note For POST endpoints that require CSRF + session auth (login), the skill should: 1. Fetch login page to get CSRF token 2. POST login with credentials 3. Use cookie jar for all subsequent requests ``` ### Technical Analysis Plaintext HTTP on a loopback interface does not inherently send information to an external network. However, the Skill accepts an arbitrary base URL from `INTERCEPT_BASE_URL` without requiring HTTPS, validating the destination, or limiting plaintext HTTP to loopback addresses. Consequently, an environment value such as `http://remote-host:5050` would cause the Agent to submit the username, password, CSRF token, and subsequent session cookie over an unencrypted network connection. API responses may also contain sensitive signal intelligence, observer coordinates, GPS positions, wireless-device information, recordings, decoded communications, system diagnostics, and remote-agent information. The instruction to maintain a cookie jar increases exposure if the jar is written to a predictable or persistent location. The reviewed instructions do not specify restrictive file permissions, ephemeral storage, deletion after use, or protections against redirects to another origin. The flagged login-page request is functionally necessary for the server's CSRF authentication flow. It is ...[truncated 2004 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill documents logging in with default admin credentials ('admin/admin') and provides operational steps to obtain CSRF tokens and session cookies. Publishing and operationalizing default credentials materially lowers the barrier to unauthorized access, and in this skill's context that access enables surveillance, data retrieval, process control, and remote-agent management.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The documentation exposes destructive or system-affecting operations such as POST /killall without any warning, approval requirement, or operational safeguards. In the context of an agent skill, this can translate a natural-language request into service disruption or termination of active SDR processes, making accidental or unauthorized impact much more likely.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill description is extremely broad and encourages activation for almost any SDR, SIGINT, scanning, or remote-control task. In an agent setting, that increases the chance the skill is invoked for sensitive or risky actions without strong user confirmation or scope checks, especially because the documented API includes surveillance, network scanning, process control, and remote-agent operations.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

This markdown file describes endpoints that expose detailed device debug information, including USB topology, permissions, and driver state, but provides no warning about the sensitivity of that information. The same omission recurs for other endpoints that start radio collection, stream aircraft/ACARS/APRS data, or modify Wi‑Fi interface state, which could affect privacy or system integrity.

Content

No source excerpt is available for this finding.