testsmash
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection from the codebases it audits.
- Ingestion points: The coordinator and fanned-out agents ingest the entire codebase, including tests, harnesses, and eval configurations in SKILL.md.
- Boundary markers: No specific boundary markers or instructions to ignore embedded content are provided for the data processed from files.
- Capability inventory: The skill is authorized to execute shell commands to run tests and utilize mutation probes which involve code manipulation.
- Sanitization: No sanitization of ingested code content is performed before processing.
- [COMMAND_EXECUTION]: The skill instructions permit the execution of repository test suites and other shell commands. This capability is necessary for the audit but could be exploited if the agent encounters malicious instructions within the audited files.
- [DYNAMIC_EXECUTION]: The process of mutation probing involves programmatically altering code and executing the results to measure test coverage and effectiveness as described in the 'Verdicts' section of SKILL.md. This runtime code modification and execution cycle presents a potential risk if the audit targets an adversarial repository.
Audit Metadata