skills/agentskillexchange/skills/files-to-prompt Directory Concatenator for LLM Context/Gen Agent Trust Hub
files-to-prompt Directory Concatenator for LLM Context
Pass
Audited by Gen Agent Trust Hub on Jun 4, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill describes installing the 'files-to-prompt' package using 'pip install'. This is a standard dependency installation for a utility hosted on GitHub.
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data into the LLM context, which presents a surface for indirect prompt injection. However, it includes features for structural boundaries such as Claude XML formatting and Markdown code blocks. 1. Ingestion points: Reads local files and directories as specified in SKILL.md. 2. Boundary markers: Supports XML tags and Markdown code blocks to delimit file content. 3. Capability inventory: The skill focuses on file reading and concatenation tasks via the provided CLI tool. 4. Sanitization: Employs file extension filtering and gitignore-aware exclusions to limit the scope of ingested data.
Audit Metadata