The Agent Skills Directory

COMMAND_EXECUTION (SAFE): The provided bash script find-polluter.sh is a utility for locating test pollution by executing npm test on individual files. This is standard behavior for development tools.
DATA_EXFILTRATION (SAFE): Diagnostic examples in references/four-phases-detailed.md show how to inspect environment variables and keychain identities. These are legitimate troubleshooting steps for complex build/signing issues and do not involve unauthorized data transmission.
PROMPT_INJECTION (SAFE): The skill processes external error logs, which is an inherent part of its debugging purpose.
Ingestion points: Phase 1 (Root Cause Investigation) involves reading error messages and stack traces.
Boundary markers: The procedural framework of the skill acts as a constraint on how the data is interpreted.
Capability inventory: Local execution of tests via npm test and diagnostic shell commands.
Sanitization: Not applicable as the skill focuses on technical diagnostics.

systematic-debugging