clarifying-test-basis
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external materials for software testing purposes, creating an attack surface where embedded instructions in those materials could influence the agent's behavior.
- Ingestion points: In the 'Exploration Loop' section of
SKILL.md, the skill is instructed to aggregate information from PRDs, business rules, prototypes, interface contracts, architecture, code, configurations, data, logs, and historical defects. - Boundary markers: The instructions do not define specific delimiters or 'ignore' commands to help the agent distinguish between data to be analyzed and instructions to be followed within these ingested files.
- Capability inventory: The skill allows the agent to investigate facts ('能自行取得的事实先调查') and read reference files (e.g.,
references/guideword-cards.md), which implies the capability to read and process local file content. - Sanitization: There are no instructions for sanitizing, escaping, or validating the external content before it is processed or presented to the user.
Audit Metadata