The Agent Skills Directory

[SAFE]: The skill is primarily instructional and focused on documentation. It does not perform sensitive operations, execute arbitrary code from untrusted sources, or access credentials. The recommended testing tools (e.g., Playwright, Vitest) and the verification commands (e.g., grep) are standard, benign utilities for project analysis and content validation.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an inherent surface for indirect prompt injection as it is designed to ingest local project context to inform its strategy generation.
Ingestion points: The skill reads from .agents/qa-project-context.md (referenced in SKILL.md) to gather business and technical context.
Boundary markers: The instructions do not specify delimiters or warnings to treat the content of the context file as untrusted data.
Capability inventory: The skill's primary function is generating markdown text; it possesses no capabilities for system modification, privileged execution, or network exfiltration.
Sanitization: No data sanitization or schema validation is applied to the ingested context file content.

test-strategy