The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external requirement files to generate architecture designs.
Ingestion points: Reads from docs/02-requirements/SRS.md, backlog/US-*.md, and UC-*.md as specified in the review process of SKILL.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within these requirement files.
Capability inventory: The skill instructs the agent to execute shell commands via plantuml for rendering and gh for issue tracking.
Sanitization: There is no evidence of sanitization or validation of the content read from requirement files before it influences architectural decisions or command parameters.
[EXTERNAL_DOWNLOADS]: The skill references external resources for diagram rendering.
Evidence: The file templates/c4.puml includes a remote PlantUML library: https://raw.githubusercontent.com/plantuml-stdlib/C4-PlantUML/master/C4_Context.puml.
Context: This is a well-known community standard library for C4 diagrams hosted on GitHub.
[COMMAND_EXECUTION]: The skill instructs the agent to use local command-line interfaces for project tasks.
Evidence: SKILL.md directs the agent to run plantuml -tsvg to render diagrams. templates/issue-tracker-github.md provides instructions for using the gh CLI for managing GitHub issues, including gh issue create and gh issue view commands.

ooad-architect