The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted RSS content (titles, snippets, and full article text) fetched from external Miniflux feeds. This presents a potential surface for Indirect Prompt Injection where malicious instructions embedded in a news article could attempt to influence the agent's behavior during summarization.
Ingestion points: Fetches data from external RSS feeds via the Miniflux API in scripts/miniflux.py.
Boundary markers: The agent instructions in SKILL.md lack explicit delimiters (e.g., XML tags) to separate untrusted content from the system prompt.
Capability inventory: The skill can perform network requests and mark entries as read.
Sanitization: The script includes a _HTMLStripper class to remove HTML tags, but it does not sanitize the text content for potential prompt injection patterns.
[DATA_EXPOSURE]: The skill manages Miniflux API credentials (URL and Token) by storing them in a local configuration file at ~/.config/clawdbot/miniflux-news.json. The implementation follows security best practices by applying restrictive file permissions (0600) during the configuration process in the cmd_configure function.

miniflux-news