The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data fetched from external URLs defined in sources.json.
Ingestion points: External content is ingested via WebFetch and mcp__chrome-devtools__* tools during Phase 2, 3, and 6.
Boundary markers: Absent. The instructions for sub-agents (Phase 3 and SubAgent calling examples) do not employ explicit delimiters or instructions to ignore embedded commands within the fetched HTML or text.
Capability inventory: The skill possesses significant capabilities including Write (to create reports and update cache.json), Bash (for directory and file management), and Task (for multi-agent orchestration).
Sanitization: Absent. There is no evidence of content sanitization or escaping before the external data is passed to the model for summarization and quality scoring.
[EXTERNAL_DOWNLOADS]: Fetches data from well-known technical and productivity services including Hacker News, HuggingFace, Product Hunt, and various industry blogs. These network operations are consistent with the skill's primary purpose of news aggregation.
[COMMAND_EXECUTION]: Utilizes the Bash tool for local environment operations including mkdir, date, and ls. These commands are restricted to specific patterns required for directory creation and timestamping of generated reports.

daily-news-report