The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill represents an indirect prompt injection surface as it ingests and processes untrusted external data (Gherkin .feature files and Cucumber JSON reports) to perform its core functions of inventorying and refactoring steps.
Ingestion points: External .feature files and project JSON reports (e.g., reports/cucumber.json) are processed to extract step patterns as detailed in references/coverage-map.md and references/step-extraction-and-overlap.md.
Boundary markers: The instructions do not include specific delimiters or "ignore embedded instructions" markers when reading the content of these files.
Capability inventory: The skill uses sed to rewrite source files, grep for pattern extraction, and git for CI checks. These capabilities are used for the intended purpose of code refactoring and library maintenance.
Sanitization: The provided scripts/step-overlap.py implements sanitization by normalizing strings (converting to lowercase, removing articles, and collapsing whitespace) before patterns are analyzed, which mitigates simple injection attempts within step texts.

bdd-step-library-curator