The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes several Python and Bash scripts that execute shell commands using subprocess.run(). These commands are used for managing the lifecycle of GECX agents, such as pulling/pushing configurations via the cxas CLI, running linting tools, and executing test suites. This behavior is consistent with the skill's primary purpose as a developer orchestration toolkit.\n- [INDIRECT_PROMPT_INJECTION]: The skill includes analysis tools (coverage-analyst, triage-failure) that ingest data from agent instructions and evaluation results. This data is processed using LLMs (Gemini) to generate reports and diagnoses. While this creates a surface for indirect prompt injection if malicious instructions are embedded in the source files, the skill's utility is focused on developer workflows within a trusted environment.\n- [EXTERNAL_DOWNLOADS]: The setup.sh script installs the cxas-scrapi package from local source and dependencies via pip/uv. This is a standard setup procedure for Python-based developer tools.\n- [DYNAMIC_EXECUTION]: The scripts/tests/test_cluster_failures.py uses __import__ to dynamically load the module under test. This is a common and safe pattern in unit testing.

cxas-agent-foundry