feedback-generator

Warn

Audited by Gen Agent Trust Hub on Aug 25, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill contains a shell script block intended to submit feedback logs via a CLI tool. It uses python3 -m cli feedback submit with arguments like --domain and --conversation-id that are derived from user session data. This presents a risk of command injection if the agent interpolates unsanitized user-provided strings into the shell execution environment.
  • [EXTERNAL_DOWNLOADS]: The instructions suggest the installation of an external package vidbyte-skills via npm install -g vidbyte-skills if the CLI is not present. This encourages the user or agent to download and install third-party code from a public registry.
  • [DATA_EXPOSURE]: In 'Passive Mode', the skill automatically writes session feedback logs to a local skills directory. While intended for logging, this behavior involves writing data inferred from user conversations to the local file system without explicit per-action confirmation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 25, 2026, 09:14 PM
Security Audit — agent-trust-hub — feedback-generator