clawteam-dev
Pass
Audited by Gen Agent Trust Hub on Mar 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs recursive directory deletion (
rm -rf) on subdirectories within the~/.clawteam/directory to reset environment state. It also manages system processes and git state usingtmux kill-sessionandgit worktree remove --force.- [DATA_EXFILTRATION]: The skill enumerates local directories, git branches, and worktrees usinggit worktree list,git branch, andls ~/.clawteam/. While no external network exfiltration was detected, the skill accesses and verifies tool-specific data structures on the local file system.- [PROMPT_INJECTION]: The skill ingests and displays messages from external agents viaclawteam task waitandclawteam inbox receive. This presents an indirect prompt injection surface where malicious output from a worker agent could attempt to influence the supervisor agent. Remediation: Wrap external agent content in explicit delimiters (e.g., XML tags) and include instructions for the supervisor to ignore any embedded commands or directives within those markers.
Audit Metadata