subagent-driven-development
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process implementation plans and task lists which may originate from untrusted external sources (e.g., repository files, PR descriptions, or user-provided documents).
- Ingestion points: The workflow involves reading an implementation plan and extracting tasks with full text (SKILL.md).
- Boundary markers: There are no specified delimiters or instructions to prevent the agent from following malicious instructions embedded within the implementation plans or tasks.
- Capability inventory: The subagents are granted capabilities to implement code, run tests, and commit changes, which involves file system writes and shell command execution (SKILL.md).
- Sanitization: The process does not describe any validation, escaping, or filtering of the external plan content before it is processed by the subagents.
- [EXTERNAL_DOWNLOADS]: The skill metadata references an external repository located at
github.com/sickn33/antigravity-awesome-skills. No automatic download or execution of scripts from this source was detected in the provided skill instructions.
Audit Metadata