setup-pstack
Pass
Audited by Gen Agent Trust Hub on Nov 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to manage local configuration files (~/.claude/rules/pstack-models.md or ~/.agents/pstack-models.md) to define model preferences for different roles.
- [COMMAND_EXECUTION]: Executes standard host-provided commands (pi --list-models, codex exec -m) to discover available models. These are used for legitimate environment discovery and do not involve user-supplied command injection.
- [DATA_EXPOSURE]: Reads the host-adapter.md and pstack-models.md files. These are local configuration files related to the skill's operation and do not contain sensitive user credentials, SSH keys, or environment secrets.
- [INDIRECT_PROMPT_INJECTION]: The skill reads a local configuration file, but it implements validation steps to ensure that any chosen models are within the set of confirmed available models on the host, reducing the risk from malformed configuration content.
Audit Metadata