The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses the html2canvas library, which is loaded from Cloudflare's well-known CDN (cdnjs.cloudflare.com). The script tag includes the integrity attribute (SRI hash) and crossorigin attribute, ensuring the library's authenticity and preventing unauthorized modifications during transit.
[INDIRECT_PROMPT_INJECTION]: The skill processes technical project data (PR changes, architecture) to generate diagrams. While this is an ingestion surface for potentially untrusted data, the instructions explicitly advise against embedding raw diff content (the primary vector for such attacks) and focus on high-level summaries, which naturally mitigates the risk of the agent obeying hidden instructions in the code it analyzes.
[COMMAND_EXECUTION]: The skill references standard Git operations (git rev-parse --git-common-dir) to determine the repository root for file placement. This is a routine development action for an AI agent and does not exhibit malicious intent or privilege escalation patterns.

diagram-generation