The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill includes find-polluter.sh, a utility script that automates running npm test on specific files to find state pollution. This is a standard development task performed on local files and does not involve remote code or privilege escalation.
[PROMPT_INJECTION] (SAFE): The CREATION-LOG.md and test scenarios include 'bulletproofing' instructions that reinforce the systematic debugging process against user-induced time pressure. These are quality-of-service constraints for reliability and do not attempt to bypass safety filters or extract system prompts.
[INDIRECT_PROMPT_INJECTION] (SAFE): The skill defines a surface for processing local code and test results. 1. Ingestion points: find-polluter.sh reads file paths; condition-based-waiting utilities process thread events. 2. Boundary markers: Absent. 3. Capability inventory: Local command execution (npm test). 4. Sanitization: Variables are properly quoted in the shell script to prevent command injection. The surface is necessary for the skill's function and lacks malicious intent.

systematic-debugging