The Agent Skills Directory

[SAFE]: The skill performs legitimate system design tasks. It identifies existing infrastructure using standard directory listing commands (ls -a) to gather context for its architectural recommendations.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it ingests data from local infrastructure files (e.g., database configs, service definitions) to inform its design process.
Ingestion points: Discovery of local files via ls -a and the Read tool in Step 0 of SKILL.md.
Boundary markers: The instructions do not define delimiters or specific "ignore instructions" markers for the content read from local files.
Capability inventory: The skill is authorized to use Bash, Write, Edit, and WebFetch tools.
Sanitization: No explicit sanitization or validation of local file content is described.
[SAFE]: The metadata in both SKILL.md and plugin.json is consistent and correctly identifies the author and purpose of the skill. The unusual formatting instructions in the 'Delivery' section appear to be a template artifact and do not impact the safety of the skill.

spine-design