The Agent Skills Directory

[SAFE]: The skill operates entirely on local project data to maintain a test registry. It does not perform network operations, access sensitive system credentials, or execute arbitrary code. All file operations are restricted to project-specific paths like tests, design documentation, and quality assurance logs.
[PROMPT_INJECTION]: The skill ingests data from external files (bug reports and design documents) which could potentially contain malicious instructions. This represents an indirect prompt injection surface. However, the risk is minimal as the skill's capabilities are limited to reading file metadata and writing a specific markdown manifest, with no high-privilege tools (such as shell execution) available.
Ingestion points: Reads files from design/gdd/, production/qa/bugs/, and tests/ directories.
Boundary markers: No explicit delimiters or instructions to ignore embedded content are defined in the file parsing steps.
Capability inventory: The skill uses Read, Glob, Grep, Write, and Edit tools. It lacks Shell or network-capable tools.
Sanitization: No specific sanitization or validation of the ingested file content is mentioned before it is processed for the manifest report.

regression-suite