The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled Python script (scripts/run.py) to automate the fetching, scoring, and organization of daily news topics.
[EXTERNAL_DOWNLOADS]: The skill retrieves data from numerous external endpoints, including the Hacker News API, RSS feeds for various tech blogs (OpenAI, Hugging Face, etc.), and HTML pages from sites like TLDR AI and James Clear.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it ingests untrusted third-party data and passes it to the agent for complex reasoning tasks (translation and summarization).
Ingestion points: Articles, titles, and summaries are fetched from a wide variety of external web sources defined in config/sources.yaml.
Boundary markers: There are no clear delimiters or instructions telling the agent to treat the fetched article content purely as data or to ignore instructions embedded within those strings.
Capability inventory: The skill has access to powerful capabilities including Bash, Write, and Edit, which could be abused if an injection is successfully processed by the agent.
Sanitization: While the Python logic strips HTML tags and truncates summaries, it does not perform any sanitization specifically targeting prompt injection patterns.

daily-topic-selector