skills/dosu-ai/dosu-skill/dosu/Gen Agent Trust Hub

dosu

Pass

Audited by Gen Agent Trust Hub on Jul 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the Dosu CLI by executing a script fetched from the vendor's official GitHub repository (raw.githubusercontent.com/dosu-ai/dosu-cli/main/install.sh). It also utilizes the @dosu/cli package from the npm registry.
  • [COMMAND_EXECUTION]: Executes a variety of subcommands via the dosu CLI to perform platform operations such as dosu setup, dosu ask, dosu docs, and dosu review. It also performs codebase inspection using standard Git and file system commands during the audit workflow.
  • [DATA_EXFILTRATION]: Manages platform authentication by retrieving and storing Supabase access tokens and API keys locally. These credentials are used to authorize requests to the Dosu platform's tRPC and backend services.
  • [PROMPT_INJECTION]: The skill processes untrusted data from external sources, including documentation from platforms like Notion and Confluence, as well as conversation threads from GitHub and Slack. To mitigate risks associated with indirect prompt injection, the skill enforces safety protocols requiring the agent to display content diffs and obtain explicit user confirmation (--confirm) before approving or applying changes.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 1, 2026, 07:50 PM
Security Audit — agent-trust-hub — dosu