The Agent Skills Directory

[PROMPT_INJECTION]: Indirect Prompt Injection vulnerability surface. The skill is designed to ingest untrusted source code from the user's project to generate and potentially execute test suites.\n
Ingestion points: The skill uses Grep and Read tools to extract protocol definitions and behaviors from local files (SKILL.md, Phase 1).\n
Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following malicious instructions that might be embedded in the comments or metadata of the analyzed source code.\n
Capability inventory: The skill is scoped with Write, Edit, and Bash capabilities, which allows it to create executable test files and run shell commands based on its analysis of external code.\n
Sanitization: No sanitization or validation of the extracted protocol content is mentioned before it is interpolated into code generation templates.\n- [COMMAND_EXECUTION]: Dynamic code generation and execution. The skill automates the creation of Swift source files and suggests a workflow where these generated files are executed to verify protocol conformance. While this is the intended primary purpose of a testing skill, it involves the runtime generation and execution of code derived from user-influenced inputs.

test-contract