Back to skill

Security audit

sn-code-video

Security checks for vulnerabilities and agentic risk

Overview

The video workflow fits its purpose, but crafted inputs can execute code or overwrite files outside the chosen output folder.

Review before running with access to important files. Validate narration IDs as simple filenames, pass contact-sheet settings as validated numbers rather than executable source, and enforce real directory containment in the capture server. Until fixed, use an isolated environment with trusted project files and arguments. Narration goes to Microsoft's speech service; use only text approved for external processing. Dependency versions and full runtime behavior were not verified.

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
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
Findings (12)

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 78)May include surrounding context.

md
- **Browser pages:** `scripts/capture.mjs` serves the project folder, opens the page in headless Chromium (Playwright or Puppeteer, resolved from the project's

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

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.

The narration workflow sends script text to Microsoft's TTS service without explicitly disclosing that user-provided narration leaves the local system.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The workflow defaults to cloud TTS, which transmits narration text to Microsoft. Although the provider and network requirement are identified, the guide does not explicitly explain that script contents leave the system or require authorization for potentially confidential narration.

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/sync_check.py (reported line 19)May include surrounding context.

python
thr = float(args[args.index("--thr") + 1]) if "--thr" in args else 0.7   # lower it for very quiet effects
SR = 48000

raw = subprocess.run(["ffmpeg", "-v", "error", "-i", f, "-vn", "-ac", "1", "-ar", str(SR), "-f", "s16le", "-"], capture_output=True).stdout
x = np.frombuffer(raw, "<i2").astype(float) / 32768
if not len(x): sys.exit("no audio decoded")

Narration text is sent to an external speech service without explicit privacy disclosure.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The script passes narration text to edge-tts and streams the response from an external speech service without a privacy notice or explicit external-processing gate. Narration may contain confidential user material, so normal video generation can disclose that material even though the transmission serves a legitimate purpose.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
96% confidence
Finding

The subprocess uses an argument list and is not shell-injectable, but its output path incorporates an unvalidated line ID from the JSON configuration. Absolute IDs or traversal components can escape OUT_DIR, and FFmpeg's -y option permits overwriting an existing WAV outside the project; the related MP3 and metadata writes also use these unsafe paths.

Content

Scanner excerpt · scripts/tts.py (reported line 60)May include surrounding context.

python
except Exception as e:
                    if attempt == 2: raise
                    print("retry", ln["id"], e)
            subprocess.run(["ffmpeg", "-y", "-v", "error", "-i", str(mp3), "-ac", "1", "-ar", str(SR), str(wav)], check=True)
            meta.write_text(json.dumps({"key": key, "words": words}, ensure_ascii=False))
        words = json.loads(meta.read_text())["words"]
        a = load(wav)

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

The NumPy requirement permits installations to resolve to different releases and transitive dependencies without review. This is a supply-chain hardening weakness, not evidence that the installed package is compromised or vulnerable.

Content

Scanner excerpt · requirements.txt (reported line 2)May include surrounding context.

text
# system: ffmpeg (libx264, aac, libmp3lame; filters loudnorm, ebur128, silencedetect, astats)
numpy
scipy
soundfile
librosa          # beat tracking and audio features for music-driven films

Unverifiable Dependency: numpy has 16 known advisory(ies) (CVE-2014-1859 (Numpy arbitrary file write via symlink attack); CVE-2021-41495 (NumPy NULL Pointer Dereference); CVE-2021-33430 (NumPy Buffer Overflow (Disputed)) +13 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
40% confidence
Finding

OSV returned advisories for this package, but its resolved version is unknown. A package-name lookup does not establish whether the installed release is affected by those advisories.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

The SciPy requirement leaves the selected release uncontrolled across installations. Installing tools is part of this skill's workflow, so an unexpected or compromised future release could enter the environment without a manifest change.

Content

Scanner excerpt · requirements.txt (reported line 3)May include surrounding context.

text
# system: ffmpeg (libx264, aac, libmp3lame; filters loudnorm, ebur128, silencedetect, astats)
numpy
scipy
soundfile
librosa          # beat tracking and audio features for music-driven films
edge-tts         # default TTS (needs network access to speech.platform.bing.com)

Unverifiable Dependency: scipy has 4 known advisory(ies) (CVE-2013-4251 (SciPy creates insecure temporary directories); CVE-2013-4251 (The scipy.weave component in SciPy before 0.12.1 creates insecure temporary dire); CVE-2023-25399 (A refcounting issue which leads to potential memory leak was discovered in scipy) +1 more), but the manifest does not pin a version, so it is unknown whether the installed release is affected

Low
Category
Supply Chain
Confidence
40% confidence
Finding

OSV returned advisories for this package, but its resolved version is unknown. A package-name lookup does not establish whether the installed release is affected by those advisories.

Content

No source excerpt is available for this finding.

Unpinned Dependencies

Low
Category
Supply Chain
Confidence
90% confidence
Finding

The soundfile requirement does not constrain the release or verify its artifact. This weakens reproducibility and supply-chain control for a dependency used to process audio, although no affected installed version is demonstrated.

Content

Scanner excerpt · requirements.txt (reported line 4)May include surrounding context.

text
# system: ffmpeg (libx264, aac, libmp3lame; filters loudnorm, ebur128, silencedetect, astats)
numpy
scipy
soundfile
librosa          # beat tracking and audio features for music-driven films
edge-tts         # default TTS (needs network access to speech.platform.bing.com)
faster-whisper   # narration check; pre-download models: Systran/faster-whisper-base.en, Systran/faster-whisper-small