signals-scout-feature-flags
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, such as command injection, data exfiltration, or obfuscation, were identified. The skill is highly focused on its stated purpose of feature flag analytics and health monitoring using provided MCP tools.
- [INDIRECT_PROMPT_INJECTION]: The skill explicitly recognizes the risk of indirect prompt injection from untrusted sources, such as event-supplied feature flag keys and responses, as well as health issue payloads. It implements proactive defense mechanisms by instructing the agent to treat these values as raw data only, use sanitization techniques like truncation and quoting, and strictly prohibits their inclusion in report fields that act as prompts for autonomous downstream tasks.
- [COMMAND_EXECUTION]: While the skill involves executing SQL queries and code searches via tools like 'execute-sql' and 'gh search', these operations are performed within a restricted environment using established tool interfaces for legitimate diagnostic purposes. The instructions ensure that dynamic values like flag keys are handled as string literals to prevent injection within these tool calls.
Audit Metadata