linkedin-ghostwriting
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user data (stories, insights, and results) during the Phase 1 interview process which is then interpolated into prompts for generating hooks and post bodies. This creates a vulnerability surface for indirect prompt injection.\n
- Ingestion points: User input provided during the Phase 1 Strategic Interview, Phase 2 Hook Selection, and Phase 3 Post Body generation in SKILL.md.\n
- Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are used when processing user-provided stories or insights.\n
- Capability inventory: The skill has access to file system tools (Read, Edit, Write, Glob, Grep) and the ability to delegate to other agents (Agent tool).\n
- Sanitization: No explicit sanitization or validation of user-provided content is performed before interpolation into the writing workflow.
- [SAFE]: The skill's external references and metadata are consistent with the vendor infrastructure. The use of Unicode bold characters is intended for LinkedIn formatting and does not serve as a mechanism for obfuscating instructions. No evidence of data exfiltration, hardcoded credentials, or unauthorized remote code execution was found.
Audit Metadata