The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is configured to execute local Python scripts and unit tests to facilitate practical certification labs.
Evidence: The agent is instructed to run python3 <lesson-path>/code/main.py, python3 -m unittest discover -s <lesson-path>/code/tests -v, and python3 scripts/check_mcpa_wire.py in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill fetches certification program data, lesson content, and track configurations from the author's GitHub repository.
Evidence: The skill reads files from https://raw.githubusercontent.com/rohitg00/ai-engineering-from-scratch/main/ as defined in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes certification content and quiz data from external sources, which constitutes a vulnerability surface if the content contains adversarial instructions.
Ingestion points: Fetches program.json, mcpa-f.json, en.md, and quiz.json from GitHub as specified in SKILL.md.
Boundary markers: Absent. There are no instructions for the agent to use specific delimiters or to disregard instructions found within the fetched content.
Capability inventory: The skill has the capability to execute shell commands via Python to run labs and validators.
Sanitization: Absent. The instructions do not specify any validation or filtering of the remote content before processing or interpolation into the agent's context.

mcpa-certification