The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill documentation and reference files (e.g., SKILL.md, references/prompt-migration.md) explicitly instruct the user to handle sensitive API keys via environment variables or .env files, providing a specific warning against sharing keys directly in the chat interface.
[EXTERNAL_DOWNLOADS]: The skill fetches documentation, JSON schema files, and integration guides from the official langfuse.com and github.com domains. These are functional requirements for providing current information and tool support.
[COMMAND_EXECUTION]: The skill utilizes standard command-line tools including npx, bunx, and the gh CLI. It also employs shell utilities like grep and xargs to manage local configuration settings.
[DATA_EXFILTRATION]: Functional API operations are performed against the Langfuse platform using user-provided credentials. A diagnostic step is included to verify authentication before continuing with data manipulation.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes application source code and prompt templates to facilitate migration to Langfuse.
Ingestion points: Reads local codebase files and existing prompts (references/prompt-migration.md).
Boundary markers: None identified in the provided instructions.
Capability inventory: Uses Bash (curl, npx, gh) and WebFetch tools.
Sanitization: Not explicitly defined; relies on standard agent parsing of code content.

langfuse