The Agent Skills Directory

[PERSISTENCE]: The file references/day1_onboarding.md includes instructions for users or agents to modify shell configuration files (~/.bashrc or ~/.zshrc) to persist changes to the system PATH. While common for developer environment setup, modifying shell initialization scripts is a standard persistence mechanism.
[INDIRECT_PROMPT_INJECTION]: The skill features an autonomous execution mode described in references/master_qa_prompt.md. This mode instructs the agent to read test case specifications from external Markdown files and execute the contained steps verbatim. This creates an attack surface where a malicious or compromised test case document could induce the agent to perform unauthorized operations.
Ingestion points: The agent reads from tests/docs/[CATEGORY]-TEST-CASES.md (e.g., 02-CLI-TEST-CASES.md) and tracking CSVs.
Boundary markers: None identified in the prompt instructions; there are no clear delimiters to distinguish test data from executable instructions.
Capability inventory: The agent is empowered to execute shell commands (e.g., ccpm, pnpm, docker), perform file system operations, and update tracking spreadsheets.
Sanitization: The skill does not mention any validation or sanitization of test steps before they are executed in the autonomous loop.

qa-expert