The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents and facilitates the execution of a local shell script, scripts/coverage/coverage.sh, to perform code coverage analysis. This includes running standard development tools like xcodebuild, swift test, xccov, and llvm-cov. It also includes maintenance commands for cleaning build directories using rm -rf as part of the clean build and coverage process.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests and summarizes terminal output produced by the coverage script, which constitutes a potential attack surface for indirect prompt injection from test metadata.\n
Ingestion points: The agent reads terminal output from scripts/coverage/coverage.sh (as defined in SKILL.md), which includes module names, file paths, and test metrics.\n
Boundary markers: The skill instructions do not define specific delimiters or provide safety warnings to the agent regarding the handling of the script's output as untrusted data.\n
Capability inventory: The skill enables shell execution and interacts with local build artifacts located in directories such as .xcresult and .build.\n
Sanitization: No sanitization or validation of the metrics, file paths, or identifiers is specified before the agent processes and reports the resulting data.

coverage-report