The Agent Skills Directory

[PROMPT_INJECTION]: The skill includes authoritative behavioral constraints ("The Iron Law", "Violating the letter of this process is violating the spirit") and explicitly instructs the agent to disregard user or manager pressure for "quick fixes" that skip the investigation phase. While intended to ensure technical rigor, these are patterns of instructional overrides.
[COMMAND_EXECUTION]: The framework mandates the creation and execution of diagnostic scripts and automated tests. It provides shell command examples for inspecting environment variables (env) and system keychain identities (security list-keychains) to troubleshoot build and signing issues.
[DATA_EXFILTRATION]: The skill involves accessing sensitive local diagnostic data (environment variables and keychain status). However, analysis of the instructions and examples shows no patterns of data exfiltration or transmission to external destinations.
[REMOTE_CODE_EXECUTION]: The skill mentions external identifiers such as superpowers:root-cause-tracing and superpowers:test-driven-development. These appear to be internal system references or complementary skills and do not represent untrusted remote code dependencies.

systematic-debugging