langsmith-trace

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the installation script for the LangSmith CLI from LangChain AI's official GitHub repository.\n- [REMOTE_CODE_EXECUTION]: Installs the CLI tool by executing the downloaded remote script in a shell environment.\n- [COMMAND_EXECUTION]: Configures the system shell profile (~/.zshrc) to include the installation directory in the PATH and to store authentication variables for persistence.\n- [PROMPT_INJECTION]: Retrieves and processes trace logs and execution data from the LangSmith platform, which presents a surface for indirect prompt injection if the retrieved content contains malicious instructions.\n
  • Ingestion points: Fetches run inputs and outputs from the LangSmith API via CLI commands like langsmith run get --include-io.\n
  • Boundary markers: No specific delimiters or warnings are used to isolate external trace data from the agent's core instructions.\n
  • Capability inventory: The skill has access to shell command execution and file system modification (used for tool setup).\n
  • Sanitization: No data validation or sanitization is performed on the retrieved trace content.
Recommendations
  • INFO: Downloads code from trusted source(s): https://raw.githubusercontent.com/langchain-ai/langsmith-cli/main/scripts/install.sh
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 02:20 AM
Security Audit — agent-trust-hub — langsmith-trace