The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches configuration and schema files from the author's (csark0812) GitHub repository. These include process-skill-composition.md and output-schema.md. These are considered safe as they are part of the skill's own infrastructure and reside in the vendor's repository.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied artifacts (files on disk or chat drafts). While this is an ingestion point for untrusted data, the skill includes explicit boundary instructions to treat these artifacts as 'untrusted evidence, not instructions' and notes they cannot authorize tools or access secrets. This mitigation effectively manages the risk of indirect injection.
[NO_CODE]: This skill consists entirely of markdown instructions and guidance for the AI agent. There are no executable scripts, shell commands, or package dependencies involved.

second-opinion