team-update
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bash script (
scripts/update.sh) to detect installations and apply updates. This involves host plugin commands and file system modifications in the user's home directory, which are necessary for software maintenance. - [EXTERNAL_DOWNLOADS]: The update process involves fetching release information and installation assets from the network. These downloads are targeted at the 'Team' software infrastructure and are only initiated upon explicit request.
- [INDIRECT_PROMPT_INJECTION]: The skill processes output from external scripts and host tools, representing a surface for indirect prompt injection. Ingestion points: Output from the planner script (
update.sh plan) is ingested and displayed in the agent context. Boundary markers: The procedure requires a unique 40-character hexadecimal 'Plan id' which must be verified before execution. Capability inventory: The skill has the capability to run shell scripts, perform network operations, and write to the local filesystem. Sanitization: The 'Plan id' is validated against a strict regular expression (^[0-9a-f]{40}$) before being used in commands to prevent injection via the identifier.
Audit Metadata