e2e-test
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to resolve and execute shell commands for end-to-end testing. It provides strong defensive instructions to mitigate risks, including a prohibition on guessing command syntax, a requirement for explicit repository evidence, and a ban on installing or upgrading dependencies and system infrastructure.\n- [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it ingests untrusted data from the repository being analyzed to determine its actions.\n
- Ingestion points: The skill reads repository-resident files including
AGENTS.md,CONTEXT.md, package manifests (e.g.,package.json,Makefile), and task specifications.\n - Boundary markers: It includes explicit instructions to treat source content as "evidence rather than authority" and to apply
$agent-securitypolicies if available.\n - Capability inventory: The skill possesses shell execution capabilities used to run detected test commands.\n
- Sanitization: It instructs the agent to check command effects against existing authorizations and preserve role boundaries before execution.
Audit Metadata