tech-hub-platform
Pass
Audited by Gen Agent Trust Hub on May 18, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests untrusted data from a task prompt and interpolates it into instructions for specialist subagents.
- Ingestion points:
AGENTS.mdextracts project context, tasks, and constraints from the spawn prompt to populate subagent requests. - Boundary markers: The subagent prompt templates in
AGENTS.md(Step 4) lack clear boundary markers or instructions to ignore embedded commands within the interpolated variables. - Capability inventory: The subagents coordinated by this skill have powerful capabilities, including modifying infrastructure-as-code (Terraform/CDK), managing CI/CD pipelines, and accessing cloud services (AWS/Azure/GCP).
- Sanitization: There is no evidence of sanitization or validation of the input data before it is passed to subagents.
- Mitigating Factors: The risk is significantly reduced by the skill's own internal protocols, which require 'Explicit Approval' for production deployments, credential access, and security group modifications, as well as 'Mandatory Collaborations' with a Security Lead for production work.
Audit Metadata