The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill suggests verifying configuration by executing echo $LANGSMITH_API_KEY, which prints sensitive credentials to the terminal and stores them in shell history files.
[COMMAND_EXECUTION]: The skill instructs users to modify their shell profile (~/.bashrc) to persist the LANGSMITH_API_KEY and LANGSMITH_PROJECT variables. This establishes a persistence mechanism that stores sensitive credentials in plain text in a long-lived configuration file.
[EXTERNAL_DOWNLOADS]: The skill installs the langsmith-fetch package from a repository associated with LangChain AI, a well-known and trusted technology provider.
[PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by analyzing data fetched from the LangSmith API.
Ingestion points: Data enters the agent's context through execution traces and thread logs fetched via the langsmith-fetch command (e.g., recent-traces.json).
Boundary markers: The instructions do not provide delimiters or security warnings to the agent to prevent it from obeying instructions embedded within the trace data.
Capability inventory: The agent has the capability to execute shell commands (pip, mkdir, grep), modify environment configuration, and create local directories.
Sanitization: The skill does not implement validation or sanitization of the JSON data fetched from the external API before the agent processes and analyzes it.

langsmith-fetch