channel-quality-analysis

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of natural language instructions and YAML configuration. There are no scripts (Python, JavaScript), executables, or binary files included in the skill package.
  • [COMMAND_EXECUTION]: The skill provides instructions for the agent to use an external tool (ae-cli) to perform data queries. These commands are parameter-controlled and follow a strict delegation protocol where the skill specifies the query intent and time range, and the tool handles the execution.
  • [INDIRECT_PROMPT_INJECTION]: The skill includes a 'Relay Summary Recognition' feature designed to process context from previous sessions. While this is an ingestion point for external data, the skill uses a structured template and multi-stage verification to ensure the data is relevant to channel analysis, minimizing the risk of prompt injection from data sources.
  • [DATA_EXFILTRATION]: All data operations involve querying internal project metrics using ae-cli. The skill explicitly forbids proactively mentioning specific client names or citing internal data in archived documents, and does not perform network operations to non-whitelisted or suspicious external domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 03:22 AM
Security Audit — agent-trust-hub — channel-quality-analysis