Morfeo Content Pipeline
Security checks across static analysis, malware telemetry, and agentic risk
Overview
The skill describes a coherent TikTok content pipeline, but it explicitly tells the agent to bypass Gemini moderation and run persistent real-brand draft-posting automation without clear account or approval boundaries.
Do not enable this as-is unless you control the relevant brand assets and social accounts, have reviewed the referenced runtime project, remove the moderation-bypass instruction, and add explicit approvals, account scopes, logging, and stop controls for the scheduled draft-posting process.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The agent may try to generate content that a provider’s safety system would otherwise block, increasing policy, account, and reputational risk.
The skill instructs the agent to change prompt language specifically to avoid Gemini moderation filters, which is an explicit tool-safety bypass.
"brief_text": "...", # en INGLÉS — evita filtros de moderación de Gemini en español
Remove the moderation-evasion instruction and require normal safety-compliant prompting, refusal handling, and review when provider moderation is triggered.
The pipeline could keep consuming provider credits and creating account drafts after a user expected a one-time action.
The artifact describes a persistent scheduled PM2 process that repeatedly generates content and creates Postiz drafts, but the visible instructions do not specify user confirmation, pause/stop controls, or containment.
**Frecuencia:** 4x/día — 11:00, 15:00, 19:00, 23:00 UTC **Proceso:** PM2 (`morfeo-content`) ... `ffmpeg → Postiz DRAFT`
Require explicit opt-in before enabling the schedule, document how to stop it, log each run, and require review before anything is published.
A user may not realize the skill needs or may use existing write-capable accounts to create drafts or push repository changes.
The workflow implies write access to Postiz/TikTok drafts and a GitHub repository, while the registry metadata declares no credential, env var, or config-path requirements for those account authorities.
`ffmpeg → Postiz DRAFT` ... **Para agregar marca:** subir imagen a `images/`, agregar entrada en `marcas.json`, hacer push.
Declare required accounts and credentials, limit scopes to draft-only/repo-specific permissions, and require confirmation before any account mutation.
Viewers or brand owners could be confused or object if real brands are used without permission or if the AI disclosure is not clear enough.
The final AI reveal is a mitigating disclosure, but the skill is designed to mimic organic content from real brands, which is trust-sensitive and may require authorization and clear labeling.
Cada video simula contenido orgánico de una marca argentina real, con un plot twist final que revela que fue creado con IA por Morfeo Labs.
Use only authorized brands/assets and ensure drafts include clear, reviewable AI and sponsorship/brand-affiliation disclosures before publication.
The reviewed artifact does not show what the referenced runtime process or external project code actually does.
Only SKILL.md instructions were available for review, while the instructions reference an existing project, PM2 process, external repository, and deployments whose code and provenance were not included.
Source: unknown; Homepage: none; No install spec — this is an instruction-only skill; No code files present
Inspect the referenced project directory, PM2 service, external repo, and deployment configurations before enabling the pipeline.