relay-80-100-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [DYNAMIC_EXECUTION]: The skill describes a pattern where an agent generates source code and test scripts (e.g.,
tests/my-feature.test.ts) and subsequently executes them usingnpx tsx. This constitutes dynamic script generation and execution at runtime. - [INDIRECT_PROMPT_INJECTION]: The workflow pattern uses command output (e.g., test failures and build logs) as direct input for subsequent agent repair tasks via variable interpolation (
{{steps.run-tests.output}}). This creates an indirect prompt injection surface where external or tool-generated data influences the agent's behavior. - Ingestion points: Test results and shell command outputs are captured in SKILL.md template steps.
- Boundary markers: The provided examples do not use explicit delimiters to isolate untrusted command output within the agent's prompt.
- Capability inventory: The workflow pattern includes capabilities for file modification, package installation, shell command execution, and git commits.
- Sanitization: There is no evidence of sanitization or filtering of the command outputs before they are processed by the agent.
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to download and install the
@electric-sql/pglitepackage from the npm registry, a well-known service.
Audit Metadata