test-first-bugs
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's core workflow ingests user-provided bug reports and passes them directly to subagents via the Task tool. This creates a potential surface for indirect prompt injection where malicious instructions hidden in a bug report could influence the subagent's execution.\n
- Ingestion points: User reports are collected in Phase 1 (SKILL.md) and passed to subagents in Phase 2.\n
- Boundary markers: The skill does not define specific delimiters or instructions to ignore embedded commands within the user-provided bug description when constructing the subagent prompt.\n
- Capability inventory: The skill uses the Task tool to create subagents and executes shell commands for testing (e.g., find, grep, pytest, npm test), providing an exploitable surface if a subagent is compromised.\n
- Sanitization: There is no evidence of input validation or sanitization for the bug descriptions processed by the agent.
Audit Metadata