The Agent Skills Directory

[SAFE]: The skill is primarily focused on instructional guidance for Markdown formatting and metadata management. The provided templates and reference files follow industry best practices for Obsidian and GitHub-Flavored Markdown compatibility.
[PROMPT_INJECTION]: The skill is designed to process untrusted external data such as transcripts and notes, which is a standard surface for indirect prompt injection. The risk is mitigated by specific instructions to isolate raw data within code blocks.
Ingestion points: External source materials (transcripts, summaries, meeting notes) processed by the agent in SKILL.md.
Boundary markers: The 'Transcript Appendix' requirement enforces the use of fenced code blocks to isolate raw input from the rest of the document, although no explicit 'ignore embedded instructions' prompt is present for the initial processing step.
Capability inventory: The skill uses Read, Grep, Edit, and Write tools to interact with the local filesystem.
Sanitization: No explicit sanitization or input validation is performed on the source material.
[COMMAND_EXECUTION]: The skill demonstrates safe usage of filesystem tools. Examples provided in the documentation (such as directory creation) are for illustrative purposes and do not automate dangerous operations.
[EXTERNAL_DOWNLOADS]: All external links referenced in the documentation and examples target well-known and trusted services such as GitHub, Stripe, and official documentation sites.

vault-scribe