why

Pass

Audited by Gen Agent Trust Hub on Nov 8, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests historical data from potentially untrusted sources like git commit messages and PR comments. \n
  • Ingestion points: Git logs, GitHub CLI outputs (issues, PRs), and external integrations via MCP servers (Slack, Notion, Sentry). \n
  • Boundary markers: The prompt templates utilize curly-brace placeholders but do not define explicit escaping or delimiters for the ingested data. \n
  • Capability inventory: The skill is strictly limited to reading and synthesis. It explicitly states that it never edits code and utilizes a read-only subagent for exploration. \n
  • Sanitization: No specific sanitization or filtering logic for remote content is described, although the instructions require preserving original wording and verbatim quotes. \n- [COMMAND_EXECUTION]: The skill procedures include the execution of standard repository tools like git and the GitHub CLI. These are used appropriately for archaeology and data retrieval within the project context.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 8, 2026, 07:22 PM