changelog

Pass

Audited by Gen Agent Trust Hub on Mar 30, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface.
  • Ingestion points: Pull Request descriptions and linked issue details fetched from GitHub via gh cli (SKILL.md).
  • Boundary markers: No explicit delimiters or instructions to ignore potential embedded commands within the fetched PR content are defined.
  • Capability inventory: The skill uses gh cli to read repository data and includes instructions for curl to transmit data externally.
  • Sanitization: The instructions do not specify any sanitization or validation of the text content extracted from GitHub before processing.
  • [DATA_EXFILTRATION]: External data transmission documentation. The skill provides a curl code snippet for posting generated changelogs to a Discord webhook URL. This involves sending summarized repository activity, developer names, and issue context to a third-party service.
  • [COMMAND_EXECUTION]: Shell command usage. The skill utilizes the gh command-line tool for analyzing repository history and suggests the use of curl for Discord integration. These tools are used legitimately within the context of the skill's stated purpose.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 30, 2026, 01:20 AM