The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill clones the source code from the author's official GitHub repository (https://github.com/rheinmir/setup.git). This is an appropriate and expected action for a development-focused skill.
[COMMAND_EXECUTION]: The skill executes local Python and Bash scripts (e.g., fdk-gate.py, medic.py, gen-converters.py) to perform validation and automation tasks. These commands are scoped to the development of the framework and are used for integrity checks and UAT.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content such as YAML policies and Markdown wiki files, which could theoretically contain malicious instructions. However, the skill employs structured validation and gate checks to mitigate risks associated with malformed or malicious data.
Ingestion points: harness/poc-vendor-neutral/policy.yaml and documentation files in the wiki/ directories.
Boundary markers: Relies on the fdk-gate.py and llmwiki-validate.py scripts to enforce structure and rules.
Capability inventory: Includes shell execution for testing, linting, and synchronization scripts.
Sanitization: Validation logic in llmwiki-validate.py and fdk-gate.py checks for schema compliance and broken links.

dym-setup-contributor