The Agent Skills Directory

[SAFE]: No executable code, scripts, or binary files are present in the skill; it operates purely as a set of instructions for the AI agent within a Markdown-based framework.
[SAFE]: The skill does not perform any network operations, data exfiltration, or access sensitive system files (e.g., SSH keys or environment variables).
[SAFE]: No hardcoded credentials, tokens, or secrets were detected.
[SAFE]: The skill uses standard Markdown for all logic and templates, with no evidence of obfuscation, hidden characters, or encoding techniques.
[PROMPT_INJECTION]: Evaluation of Indirect Prompt Injection surface:
Ingestion points: User-provided text pasted into the /trend workflow (referenced in SKILL.md and references/trend_research.md).
Boundary markers: Absent; user content is processed directly to extract patterns.
Capability inventory: The skill is restricted to text generation and local file manipulation (writing to brand_voice.md via the /voice-setup workflow). No shell access, code execution, or network tools are utilized.
Sanitization: Absent.
Conclusion: The attack surface is minimal and considered safe because the skill lacks the capabilities required to perform dangerous actions even if an injection were successful.

creator-studio