cloud-weaver-teardown

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill downloads a helper script (teardown.py) from the vendor's official domain (cloudweaver.fagnerlopes.dev). This is a legitimate resource used to update repository workflows.
  • [COMMAND_EXECUTION]: Uses GitHub CLI (gh) to manage workflows, monitor runs, and delete repositories. These operations are gated by explicit user confirmation.
  • [DATA_EXPOSURE]: Performs cleanup of local configuration files and SSH keys stored in $HOME/.ssh/ and the home directory. The deletions are restricted to files following specific naming patterns related to the tool's session and identity management.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided repository names which are interpolated into shell commands. While this presents a potential injection surface, the risk is mitigated by requiring explicit user confirmation before executing any destructive actions.
  • Ingestion points: User input for the repository name captured in SKILL.md.
  • Boundary markers: Requires the user to type "sim" to confirm destructive actions.
  • Capability inventory: Includes file deletion (rm), network downloads (curl), and repository management via GitHub CLI.
  • Sanitization: No specific shell escaping is implemented for the repository name variable, though usage is within controlled automation flows.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 04:36 AM
Security Audit — agent-trust-hub — cloud-weaver-teardown