game-qa
Fail
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions in SKILL.md and references/qa-contract.md direct the agent to read and execute an "authoritative command" from the
verify.commandfield inqa/verification.json. This allows for arbitrary shell command execution if a project includes a malicious configuration file. - [PROMPT_INJECTION]: The skill processes untrusted data from multiple project files without adequate protection against indirect instructions, using them to drive execution and logic.
- Ingestion points: Contents of
qa/verification.json,GAME_DESIGN,PRODUCT_BRIEF.md, andBUILD_BRIEF.mdare ingested into the agent context. - Boundary markers: Absent; the instructions do not specify using delimiters or warnings to disregard instructions within the processed data.
- Capability inventory: Shell command execution (via the "运行一次权威 verify" directive).
- Sanitization: Absent; the skill does not include steps to validate the command string or sanitize external content.
Recommendations
- AI detected serious security threats
Audit Metadata