The Agent Skills Directory

[SAFE]: The skill defines a set of operational procedures for verifying software development tasks. It uses tools like Bash and Grep to execute local commands (e.g., npm test, tsc, grep) and parse their output for status codes and success markers.
[SAFE]: No evidence of data exfiltration, credential theft, or unauthorized network communication was found. All network-related patterns (e.g., curl to localhost) target local development environments.
[SAFE]: The instructions do not attempt to bypass AI safety filters or override system instructions. Instead, they reinforce honesty and accuracy in the agent's reporting of task status.

verification-before-completion