The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions provide multiple methods for downloading the blogwatcher-cli binary directly from a GitHub repository (github.com/JulienTant/blogwatcher-cli). This includes piping curl output into tar for extraction into system directories.
[COMMAND_EXECUTION]: The skill relies on executing the blogwatcher-cli command-line tool with various arguments, including user-provided URLs and file paths for OPML imports.
[PRIVILEGE_ESCALATION]: The installation instructions suggest extracting the downloaded binary to /usr/local/bin, a directory that typically requires administrative privileges for write access.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display content from external RSS/Atom feeds and HTML pages (via scraping).
Ingestion points: External URLs fetched via blogwatcher-cli add and blogwatcher-cli scan in SKILL.md.
Boundary markers: Absent. The example output indicates that article titles, blog names, and categories are displayed without explicit delimiters or warnings to ignore embedded instructions.
Capability inventory: Execution of shell commands via blogwatcher-cli as described in SKILL.md.
Sanitization: None specified for the content retrieved from external feeds before it is presented to the agent.

blogwatcher