coder-templates
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-defined requirements for infrastructure provisioning and interpolates this data into Terraform configurations and README documentation.
- Ingestion points: User prompts describing target platforms, compute types, and specific software needs (e.g., 'Node 20 + Postgres 16') processed according to instructions in SKILL.md.
- Boundary markers: The skill requires the agent to summarize planned changes for user approval before proceeding and to request clarification for ambiguous inputs.
- Capability inventory: The skill utilizes file-writing capabilities and execution of shell scripts and CLI tools across its operations.
- Sanitization: No explicit automated sanitization or escaping mechanisms for user input are described for the generation of Terraform blocks or GFM alerts.
- [COMMAND_EXECUTION]: The skill invokes various local shell scripts and industry-standard command-line tools to manage the lifecycle of workspace templates.
- Evidence: Executed commands include './scripts/new_template.sh', './scripts/terraform_validate.sh', 'coder templates push', 'bun run fmt', and 'bun run shellcheck'.
Audit Metadata