known-content-archivelisted

Retrieve and archive explicit URLs, local files, URL lists, or precisely named collections as immutable local source packages. Use when the user already knows what should be preserved; do not use for open-ended discovery, recommendations, account crawling, or private collections without current-turn authorization.
qihangzhang-272/agent-skill-library · ★ 1 · AI & Automation · score 62

Install: claude install-skill qihangzhang-272/agent-skill-library

# Known Content Archive Turn material the user has already identified into a bounded, auditable local archive. The current Host performs retrieval; this Skill defines the handoff and owns durable preservation. ## Boundary - Accept explicit URLs, local files, a supplied URL list, or a collection named precisely enough to enumerate. - Do not expand the scope with keyword search, related links, account history, bookmarks, recommendations, or model memory. - Use the current Mode's installed retrieval route. Prefer Agent Reach/opencli when present; do not create another cross-platform router or duplicate login storage. - Ask for current-turn authorization before accessing logged-in or user-specific content. A public URL supplied by the user needs no extra adoption Trial. - An inaccessible item is a recorded result, not permission to substitute a different source. ## End-to-end operation ### Retrieve Read or download each supplied item with the smallest available Host capability: - ordinary pages and public platform URLs: the Host's browser, web reader, or installed retrieval Skill; - X, video, podcast, social, and creator-platform URLs: the corresponding Agent Reach/opencli route already available in the Mode; - local files: direct local read; - media: download only when requested or necessary to preserve the supplied item. Keep intermediate retrieval files inside the current Case or a temporary run directory. Do not create a second global cache. ### Record the retrieval