The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to interact with the system's pass (Password Store) utility using pass show <entry> to retrieve API keys for services like dev.to. This is a recommended security practice to avoid hardcoding secrets.
[EXTERNAL_DOWNLOADS]: The skill workflow involves fetching content from various external URLs (e.g., pycoders.com, django.news, rundown.ai) to verify submission methods and policies before drafting promotional content.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external data, including writing samples for voice profiling and content from external submission pages. While this presents a potential attack surface for indirect prompt injection, the skill includes a mandatory 'Pre-publish source audit' and strict rules against inventing claims to mitigate the risk of the agent adopting malicious instructions or fabricating details.
[COMMAND_EXECUTION]: The provided GitHub Actions workflow template (references/publish-devto.yml) utilizes npx to run well-known utilities such as mmdc (Mermaid CLI) for diagram rendering and devto-cli for article publishing.

ac-writing-blog-posts