postplus-workspace
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes the vendor's postplus CLI tool and a locally bundled Python script (scripts/workspace.py) to manage project files and state.\n
- The initialization and check commands are well-defined and operate within user-specified project directories.\n- [INDIRECT_PROMPT_INJECTION]: The skill actively reads and processes data from various markdown and YAML files within the workspace to guide agent behavior.\n
- Ingestion points: Files like project.yaml, brand.md, and product.md are read during project onboarding and strategy phases.\n
- Boundary markers: The instructions include a 'Core Contract' that mandates keeping confirmed facts distinct from unknowns and research findings, which serves as a conceptual delimiter.\n
- Capability inventory: The skill can create files and directories and execute shell commands via the workspace script and CLI.\n
- Sanitization: The Python script scripts/workspace.py performs rigorous path validation, explicitly rejecting the filesystem root, user home directory, and skill installation directory as targets.
Audit Metadata