new-workflow
Warn
Audited by Gen Agent Trust Hub on May 2, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs shell command interpolation of user-supplied variables. In Step 3, the project 'slug' and the number of rounds 'N' are directly placed into a
mkdircommand string. This lacks sanitization, allowing for potential command injection if a user or a malicious prompt provides values containing shell metacharacters (e.g., ";", "&", "|", or backticks). - [EXTERNAL_DOWNLOADS]: The skill fetches image assets from the well-known service fal.media using curl. This involves downloading files from external URLs and saving them to the local file system as part of the generation workflow.
- [PROMPT_INJECTION]: The skill ingests untrusted data from the user in the form of a project 'Brief' and 'Reference Links.' These inputs are used to build prompts for AI tasks, creating a surface for indirect prompt injection.
- Ingestion points: Metadata gathering (Step 2), brief.md template (Step 5)
- Boundary markers: None present in the prompt construction logic
- Capability inventory: mkdir, curl, jq, git, Write, Edit (SKILL.md)
- Sanitization: No sanitization or validation logic is observed for user-provided strings.
Audit Metadata