health-check
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes and formats output from the
health_checkMCP tool into a report for the user. If the underlying tool output were compromised, it could attempt to influence the agent's behavior during the reporting phase. - Ingestion points: The output of the
mcp__plugin_bitwize-music_bitwize-music-mcp__health_checktool is ingested and formatted inSKILL.md. - Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore embedded commands within the tool output.
- Capability inventory: The skill is restricted to platform tools (
ToolSearch,bitwize-music-mcp) and explicitly prohibits the use of Bash commands. - Sanitization: None identified; the skill directly renders data from the tool's result fields.
- [SAFE]: The skill uses internal platform mechanisms (MCP) and references local configuration paths (
~/.bitwize-music/venv) that are consistent with the functional purpose and the author's namespace ('bitwize-music-studio'). The explicit restriction against using Bash is a positive security constraint.
Audit Metadata