test-first
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's protocol for bug fixes requires the agent to ingest external bug reports and use them as the source of truth for creating reproduction tests. This creates a surface where a malicious bug report could contain instructions designed to influence the agent's behavior or the code it generates for testing.\n
- Ingestion points:
SKILL.md(Bug Fixes section) where bug reports are treated as the basis for reproduction tests.\n - Boundary markers: Absent. The instructions do not specify the use of delimiters or warnings to ignore embedded instructions within user-provided bug descriptions.\n
- Capability inventory:
SKILL.mdinstructions include file system deletions (The Prime Directive), shell command execution for running tests (Watch it fail), and the use of diagnosis and execution tools (godmode:fault-diagnosis,godmode:delegated-execution).\n - Sanitization: Absent. There are no instructions for sanitizing or validating the content of the bug reports before they are incorporated into the test environment.
Audit Metadata