recoup-internal-consulting-inbox-triage
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute local Python scripts including
integrations/gmail/_work/pull_threads.pyandintegrations/gmail/_work/create_draft.py. These scripts are used to query email data and create drafts within the user's Gmail environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external content from Gmail message bodies, creating an attack surface for indirect prompt injection. * Ingestion points: Untrusted email data enters the context through the
pull_threads.pyscript which reads the full body of the last message in a thread. * Boundary markers: Absent; there are no specified delimiters or instructions to treat email content as data rather than instructions. * Capability inventory: The skill has the ability to execute local scripts (subprocess calls), modify local markdown files likeAGENTS.mdandpipeline/_board.md, and interact with external integrations like Attio. * Sanitization: No automated sanitization or filtering of email content is described; however, the skill requires manual verification of all drafts and strictly prohibits automated sending.
Audit Metadata