The Agent Skills Directory

[SAFE]: The skill serves as a reference for the SpacetimeDB HTTP API, providing instructions on how to construct requests for identity creation, database operations, and SQL execution. The instructions are purely descriptive and do not initiate any automated actions or network connections.
[DATA_EXFILTRATION]: Analysis of the authorization documentation shows proper use of placeholders for bearer tokens and identity strings. The skill does not contain hardcoded credentials, nor does it provide instructions that would lead to the exposure of sensitive environment variables or local files.
[INDIRECT_PROMPT_INJECTION]: The skill documents API endpoints that accept untrusted input, specifically SQL statements and reducer arguments. It correctly identifies the attack surface by recommending the use of specialized SQL handling skills and noting the application of Row Level Security (RLS) and identity-based authorization to gate access to data.
[DYNAMIC_EXECUTION]: While the documentation mentions the publication of WebAssembly (WASM) modules to the database, this is described as a standard administrative function of the SpacetimeDB platform. The skill itself does not contain logic to generate, compile, or execute code dynamically on the agent's host system.

spacetimedb-http