credits-inventory
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes untrusted content from emails, which could contain instructions designed to manipulate the agent's behavior.
- Ingestion points: Email data is retrieved via the
gscli gmail readcommand inSKILL.md. - Boundary markers: No explicit delimiters or instructions are provided to the agent to treat email content as untrusted or to isolate it from the skill's instructions.
- Capability inventory: The skill uses shell commands (
gscli) and has file system access to write state to~/.claude/skills/credits-inventory/. - Sanitization: There is no described sanitization, validation, or filtering of the email body content before processing.
- [COMMAND_EXECUTION]: The skill relies on shell command execution via
gsclito interact with Gmail, which is a powerful capability that processes dynamic inputs like search queries and message IDs.
Audit Metadata