granola-ci-integration
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill documents a pipeline designed to process untrusted meeting note content. The provided GitHub Action template in
SKILL.mduses shell interpolation (${DECISIONS}) inside a heredoc (cat << EOF), creating a potential command injection vulnerability if meeting notes contain characters like backticks or subshell markers. This is a vulnerability surface for the system the user is instructed to build.
Audit Metadata