The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Salesforce CLI (sf apex run test) to execute unit tests within a target Salesforce organization. This is the primary intended functionality of the skill and uses standard, well-known tooling.
[PROMPT_INJECTION]: The skill includes a post-tool hook (hooks/scripts/parse-test-results.py) designed to parse output from the test execution command and generate structured instructions for an 'Agentic Test-Fix Loop'. While this mechanism processes external data (test failures) that could theoretically contain adversarial strings intended to influence the agent's next actions, the implementation is consistent with the skill's stated purpose of automating test repairs.
Ingestion points: The parse-test-results.py script ingests data via the TOOL_OUTPUT environment variable, which contains stdout from the sf CLI execution.
Boundary markers: The script does not employ specific boundary markers or 'ignore' instructions for the parsed messages in its output to the agent.
Capability inventory: The agent environment has capabilities to read and write Apex source files and execute system commands via the sf CLI.
Sanitization: The script truncates failure messages and stack traces but does not specifically filter for natural language instructions or prompt injection patterns.

running-apex-tests