The Agent Skills Directory

[CREDENTIALS_UNSAFE]: A plaintext API key is hardcoded in the 'Project Context' section of SKILL.md (value: 7a4b2ef1a7b209016ce2670b6b01fc3066a3553bfae0b0745aaf44266dba724c). This key is intended for authentication with the am_server backend and is not a placeholder.
[REMOTE_CODE_EXECUTION]: The skill facilitates remote shell command execution on Sprites microVMs via a WebSocket-based Exec API (wss://api.sprites.dev/v1/sprites/{name}/exec), as documented in both SKILL.md and references/api-workflows.md.
[COMMAND_EXECUTION]: The workflow documentation in references/api-workflows.md explicitly instructs the use of management commands such as systemctl restart, pkill, and ps aux to control services and monitor processes on remote systems.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data that could contain malicious instructions, such as tailing logs or reading files from remote microVMs. Ingestion points: Filesystem API (GET /v1/sprites/{name}/fs/{path}) and execution output from tailing logs in references/api-workflows.md. Boundary markers: Absent; the instructions do not specify delimiters or warnings to ignore embedded commands in the data. Capability inventory: Broad capabilities including remote shell access, filesystem manipulation, and service management. Sanitization: No evidence of input validation or content sanitization is provided.

sprites