The Agent Skills Directory

[SAFE]: The skill implements a navigational advisor in scripts/affordances.py that processes JSON states to recommend testing playbooks. This script uses strict enumeration validation and verifies that all file paths are relative to the skill's base directory, preventing path traversal or arbitrary file access.
[SAFE]: The provided test harness library (assets/test-harness-base/lib/harness.sh) implements defensive process management. It verifies process start times in /proc before sending termination signals to ensure process groups are only signaled if they belong to the current session, preventing accidental interference with unrelated system processes.
[SAFE]: The instructions in SKILL.md and README.md explicitly enforce security best practices, such as requiring user authorization for repository modifications, forbidding silent installations, and mandating the redaction of sensitive data from logs.
[SAFE]: Test files utilize standard testing patterns, such as subprocess.run and importlib, solely for verifying the CLI behavior and modular logic of the skill's own components.

delivery-first-testing