detect-cues

Pass

Audited by Gen Agent Trust Hub on May 7, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes local system binaries (aubioonset, aubiotrack, and aubiopitch) to analyze audio files. It constructs shell commands by interpolating user-provided inputs such as file paths and detection parameters. While the instruction template uses double quotes for the input path, it represents a standard attack surface for command injection if input is not properly sanitized by the underlying execution environment.
  • [PROMPT_INJECTION]: The skill processes untrusted external data (audio files) through CLI tools and incorporates the output into a JSON structure. This creates a surface for indirect prompt injection if subsequent skills or agent steps ingest this JSON without proper validation or sanitization.
  • Ingestion points: SKILL.md (via audio file input to CLI tools)
  • Boundary markers: Absent; the resulting JSON does not include specific boundary delimiters or instructions to ignore embedded content.
  • Capability inventory: Subprocess execution of aubio binaries and file-write operations for the JSON sidecar.
  • Sanitization: Minimal; relies on shell quoting in the command template.
Audit Metadata
Risk Level
SAFE
Analyzed
May 7, 2026, 09:13 PM
Security Audit — agent-trust-hub — detect-cues