ratel-langfuse-integrate

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted content from the user's repository (source code, manifest files, and topology data) to generate an integration plan. This content could contain instructions designed to override the agent's behavior during the generation process.
  • Ingestion points: The skill reads repository manifest files (e.g., package.json, pyproject.toml, requirements.txt) and explores source code files using the Read, Glob, Grep, and Bash tools to map the agent's architecture.
  • Boundary markers: The instructions do not specify the use of delimiters or 'ignore' warnings when interpolating file content into the agent's context during Step 1 and Step 2.
  • Capability inventory: The skill possesses the Write tool to create files in the .ratel/ directory and the Agent tool to spawn sub-tasks for exploration. It does not have network-write or code execution capabilities over the analyzed code.
  • Sanitization: There is no evidence of content sanitization or validation performed on the ingested data before it is used for plan generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 05:06 PM
Security Audit — agent-trust-hub — ratel-langfuse-integrate