context-cleanup
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines purely instructional logic for session management, providing a strategy for when and how to read configuration files (such as
.agent-team/rules/index.mdandworker.yaml) to stabilize the agent's context during phase transitions or session drift. - [COMMAND_EXECUTION]: The skill references a CLI binding
agent-team context-cleanup. This command appears to be a functional component of the agent's toolset and does not involve unsafe shell execution or injection patterns. - [DATA_EXPOSURE]: While the skill instructs the agent to read local files like
task.yamlandcontext.md, these are standard operational artifacts for the task at hand and do not target sensitive system directories, credentials, or environment files.
Audit Metadata