flaky-test-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data such as failure history, logs, and test code.
- Ingestion points: Failure history, rerun results, and test code are identified as primary inputs in
prompts/flaky-test-analysis.md. - Boundary markers: While the skill instructs the agent to "Audit the input" and "separate confirmed facts", it does not mandate specific delimiters (like XML tags) to isolate this untrusted content.
- Capability inventory: The agent is instructed to use the existing toolchain for reproduction experiments and stabilization actions.
- Sanitization: There are no explicit instructions to sanitize or filter the external content before processing.
- [DYNAMIC_EXECUTION]: The skill instructs the agent to generate and potentially execute reproduction experiments and stabilization actions.
- The
prompts/flaky-test-analysis.mdfile requires the agent to produce "concrete scenarios, assertions, verification steps, or decision gates" to stabilize tests. - Safety controls: The skill mitigates this risk by requiring root-cause claims to have reproducible evidence and advising against unauthorized production writes.
Audit Metadata