The Agent Skills Directory

[Prompt Injection] (SAFE): No instructions attempting to bypass safety filters or override system behavior were detected. The instructional content is restricted to QA methodology.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or sensitive data exfiltration patterns were found. Database connection examples use standard local development defaults.
[Unverifiable Dependencies] (SAFE): Onboarding documentation mentions standard package managers (npm, pnpm) and repository cloning for environment setup, which is consistent with the skill's purpose.
[Dynamic Execution] (SAFE): Python scripts used for metrics and validation perform local file reading and string manipulation using standard libraries. No use of eval(), exec(), or unsafe deserialization was observed.
[Indirect Prompt Injection] (LOW): The skill processes external data (CSV and Markdown).
Ingestion points: calculate_metrics.py reads a user-specified CSV path; validate_test_ids.py reads markdown and CSV paths.
Boundary markers: Absent.
Capability inventory: Limited to file reading and console output; no network or subprocess capabilities are present in the scripts themselves.
Sanitization: None; however, the impact is negligible as the scripts only perform count-based arithmetic and regex matching.

qa-expert