Back to skill

Security audit

Feed To Md

Security checks for vulnerabilities and agentic risk

Overview

This skill has a clear feed-to-Markdown purpose, but its public-only URL protection can be bypassed in a way that may let a crafted feed URL reach internal network services.

Install only if you are comfortable with a feed converter that makes outbound HTTP(S) requests from your environment. Use trusted feed URLs, avoid attacker-supplied domains, and prefer running it where internal network services or metadata endpoints are not reachable until the DNS-rebinding issue is fixed.

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 (1)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/feed_to_md.py:25
Finding

DNS Rebinding Bypasses SSRF Address Validation

Content
View full analysis
None: if hostname in {"localhost", "localhost.localdomain"}: raise ValueError(f"{label} uses localhost, which is not allowed") try: addr_info = socket.getaddrinfo(hostname, None) except socket.gaierror as exc: raise ValueError(f"Unable to resolve host: {hostname}") from exc for item in addr_info: ip_raw = item[4][0] ip = ipaddress.ip_address(ip_raw) if ( ip.is_private or ip.is_loopback or ip.is_link_local or ip.is_multicast or ip.is_reserved or ip.is_unspecified ): raise ValueError(f"{label} resolves to a non-public IP address") def validate_feed_url(raw_url: str, label: str = "Feed URL") -> str: parsed = urllib.parse.urlparse(raw_url) if parsed.scheme not in {"http", "https"}: raise ValueError(f"{label} must use http or https") if not parsed.hostname: raise ValueError(f"{label} must include a hostname") hostname = parsed.hostname.strip().lower() validate_public_hostname(hostname, f"{label} host") return parsed.geturl() ``` The validated hostname is subsequently passed to a network client that performs its own DNS resolution: ```python class PublicOnlyRedirectHandler(urllib.request.HTTPRedirectHandler): def redirect_request(self, req, fp, code, msg, headers, newurl): # noqa: D401 redirected_url = urllib.parse.urljoin(req.full_url, newurl) validate_feed_url(redirected_url, "Redirect URL") return super().redirect_request(req, fp, code, msg, headers, newurl) def fetch_xml(url: str, timeout: int = 15) -> bytes: request = urllib.request.Request( ...[truncated 3517 chars]
Remediation
View remediation
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill clearly instructs use of network access to fetch arbitrary feed URLs and can write Markdown output to files, but it does not declare any explicit tool scope such as permissions or allowed-tools. That omission weakens policy enforcement and reviewability, because an agent may invoke network and file-writing capabilities without an explicit, least-privilege declaration tied to the skill.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.