The Agent Skills Directory

[DYNAMIC_EXECUTION]: The script scripts/test_validate_agents_md.py utilizes importlib.util to dynamically load and execute the validate_agents_md.py module from the file system at runtime.
[COMMAND_EXECUTION]: The script scripts/test_skill.py invokes subprocess.run() to execute several Python scripts as sub-processes, using a path provided as a command-line argument.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading existing project files and instruction sets to inform the generation of new guidance, which represents an indirect prompt injection surface.
Ingestion points: SKILL.md directs the agent to read workspace files and existing root instruction files as evidence.
Boundary markers: No explicit delimiters or warnings are used to prevent the agent from following instructions embedded in the analyzed evidence.
Capability inventory: The skill can write to the file system and execute internal scripts.
Sanitization: There is no evidence of content sanitization for ingested data.
[EXTERNAL_DOWNLOADS]: The references/source-notes.md file contains references to documentation and resources from Anthropic, OpenAI, and Google which inform the skill's design principles.

bootstrap-agents-md