polpo
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interacts with the user's environment by executing the polpo CLI tool for project lifecycle management, including commands for logging in, linking projects, and deploying resources.
- [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of external code bundles, assets, and binary resources from GitHub repositories using the polpo skills add owner/repository command.
- [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection because it reads and processes instructions and configurations from local files that could be influenced by external contributors.
- Ingestion points: The skill reads file contents from the .polpo/ directory (such as instructions.md and agent.json) and package.json.
- Boundary markers: There are no explicit delimiters or specific instructions to isolate content read from these files from the agent's internal control logic.
- Capability inventory: The agent can invoke the polpo CLI to perform network operations, file system modifications, and deployments to cloud infrastructure.
- Sanitization: The instructions do not specify any validation or sanitization routines for the content of project files before the agent acts upon them.
Audit Metadata