Back to skill

Security audit

Podcast Agent

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent podcast generator, but its URL fetch helper can be redirected toward internal services and its external TTS data handling is under-disclosed.

Review before installing. Use this only in an environment where arbitrary article URLs cannot reach sensitive internal services, avoid feeding private or proprietary text into TTS unless you accept external processing by edge-tts/Microsoft, and prefer pinned dependencies plus hardened redirect and output-path handling.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/podcast_gen.py:28
Finding

Redirect Handling Bypasses SSRF Protection

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
SKILL.md:39
Finding

Unpinned Third-Party TTS Dependency

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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (9)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding

The skill describes and exemplifies capabilities that include network access, file writes, and shell execution, but it does not declare any explicit tool restrictions such as permissions or allowed-tools. This creates unnecessary privilege ambiguity: an agent runtime may expose broader tools than needed, increasing the risk of unintended command execution, arbitrary file output, or unrestricted web access during normal use.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The example trigger phrase is broad, conversational, and maps to a common user request, which can cause the skill to activate in contexts where the user did not explicitly intend to invoke this particular capability. Because the skill performs search, content processing, and file generation, accidental activation could lead to unnecessary network activity, content fetching, and audio creation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The voice configuration hardcodes zh-CN voices for both speakers, and the surrounding examples are oriented to Chinese output without stating that users can choose another language or locale. That can violate a language/locale policy when the skill is used in broader contexts without explicit user opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The user interaction section describes activation from ordinary natural-language requests without any narrow routing criteria or confirmation step. In an agentic environment, this can over-broaden dispatch and cause the skill to run on generic requests, leading to unintended network access, content scraping, and output generation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The TTS function sends full script text to edge-tts, which is a network-backed external service, without any explicit notice, consent, or data-handling warning. If users include sensitive article content, proprietary drafts, or personal data in scripts, that content may be transmitted off-host unexpectedly.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The manifest describes a content-to-podcast pipeline, but does not indicate any need to execute local system programs. The code checks for and runs the ffmpeg binary via subprocess, which is a stronger host capability than ordinary text processing or TTS API use and is not explicitly justified by the stated skill purpose.

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 · scripts/podcast_gen.py (reported line 224)May include surrounding context.

python
# Check if ffmpeg is available
    try:
        subprocess.run(["ffmpeg", "-version"], capture_output=True, check=True)
        # Use ffmpeg to concatenate
        subprocess.run([
            "ffmpeg", "-y",

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
79% confidence
Finding

The ffmpeg invocation itself is not shell-injectable, but it processes a concat file generated from filesystem paths while explicitly setting '-safe 0', which disables ffmpeg's path safety checks. If an attacker can influence segment or output paths, this broadens what ffmpeg may read during processing and increases risk compared with a safer concat mode.

Content

Scanner excerpt · scripts/podcast_gen.py (reported line 226)May include surrounding context.

python
try:
        subprocess.run(["ffmpeg", "-version"], capture_output=True, check=True)
        # Use ffmpeg to concatenate
        subprocess.run([
            "ffmpeg", "-y",
            "-f", "concat",
            "-safe", "0",

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
94% confidence
Finding

The function is documented as 'Merge audio segments using Python' and comments say it will 're-generate combined text,' but the implemented behavior simply reads each MP3 file and appends its bytes. This is an active mismatch between explanatory comments and actual behavior, which could mislead reviewers about what the fallback path really does.

Content

No source excerpt is available for this finding.