The Agent Skills Directory

[COMMAND_EXECUTION]: The find-polluter.sh script automates the execution of local test files using npm test. While the script employs shell quoting for variables, running tests on files discovered in the workspace involves executing local code execution pathways.
[COMMAND_EXECUTION]: The skill suggests utilizing diagnostic commands such as security list-keychains and security find-identity within SKILL.md to troubleshoot signing and build issues. These commands access system security metadata and configurations.
[INDIRECT_PROMPT_INJECTION]: The systematic debugging process requires the agent to analyze external, potentially untrusted content like error messages, stack traces, and build logs, which could contain malicious instructions.
Ingestion points: Error messages, stack traces, test results, and Git logs as described in SKILL.md.
Boundary markers: Absent; the skill does not define specific delimiters or instructions to ignore embedded commands within the ingested data.
Capability inventory: Shell command execution (npm, security, codesign, ls) and file system access across SKILL.md and find-polluter.sh.
Sanitization: Absent; the skill relies on standard shell variable quoting in helper scripts but lacks explicit prompt-level sanitization for external log content.

debug