add-provider
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data during its research phase.
- Ingestion points: OpenAPI specs, vendor source code, and live responses from
bin/gcx apiprobes (Stage 1b). - Boundary markers: The instructions do not specify any delimiters or ignore-instructions for the external content processed during discovery.
- Capability inventory: The skill enables the agent to write Go files, documentation, and execute shell commands via
miseandbin/gcx(Stage 3 and 4). - Sanitization: No automated sanitization is described for data returned from external API endpoints.
- [COMMAND_EXECUTION]: The skill orchestrates several local shell commands to manage the development lifecycle.
- Evidence: Use of
mise run gate,mise run all, and variousbin/gcxcommands for API probing and verification. - Mitigation: The skill enforces safety checks, such as requiring user verification for destructive commands and using
providers.ConfigLoaderto manage secrets safely.
Audit Metadata