create-llms

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to analyze and read files within a repository (such as README.md, documentation in /docs/, and configuration files) to generate a summary and a list of links. This represents an attack surface where malicious instructions embedded in the repository files could attempt to influence the agent's behavior during the analysis phase.
  • Ingestion points: The agent reads the repository structure, README files, and documentation directories (SKILL.md Steps 2, 3, and 5).
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the repository content analysis.
  • Capability inventory: The skill possesses the capability to write a new file (llms.txt) to the repository root.
  • Sanitization: There are no explicit instructions to sanitize or filter content extracted from the repository files before summarizing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 06:50 AM
Security Audit — agent-trust-hub — create-llms