soundcraft

Pass

Audited by Gen Agent Trust Hub on Oct 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill documents and enables extensive automation of DAW tasks through multiple interfaces, including a CLI tool (soundcraft-cli), an MCP server, and a JSON-over-TCP control channel. These are functional features designed for legitimate scripting and integration.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data formats such as audio files, MIDI files, and JSON sessions, representing an ingestion surface for untrusted data.
  • Ingestion points: Untrusted data is ingested through session opening and media import commands (crates/engine/src/io.rs).
  • Boundary markers: The documentation specifies the use of --automation-read-root and --automation-write-root flags to confine file access to specific directories (apps/soundcraft-cli/src/automation_args.rs).
  • Capability inventory: The skill includes the ability to write to the file system (saving sessions and exports) and execute local commands via the CLI and control channels.
  • Sanitization: JSON session files are validated against a schema and versioning system (crates/model/src/session.rs).
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 19, 2026, 06:30 PM