{ "schema_version": 2, "kind": "shortening-method", "format": "agent-skill", "id": "recommendations", "name": "Recommendations", "category": "Work", "summary": "Extract short, practical actions from a longer source.", "use_cases": [ "Reports", "Advice articles", "Project retrospectives" ], "word_count": 105, "url": "https://sho.rten.it/methods/recommendations/", "instructions_url": "https://sho.rten.it/methods/recommendations/SKILL.md", "skill_url": "https://sho.rten.it/methods/recommendations/SKILL.md", "json_url": "https://sho.rten.it/methods/recommendations/llms.txt", "plain_text_url": "https://sho.rten.it/methods/recommendations/prompt.txt", "license": "MIT", "sources_url": "https://sho.rten.it/sources/#recommendations", "skill_name": "recommendations", "skill_description": "Extract short, practical actions from a longer source. Use for Reports, Advice articles, Project retrospectives.", "agents_md_url": "https://sho.rten.it/methods/recommendations/AGENTS.md", "sources": [ { "title": "Fabric ยท extract_recommendations", "url": "https://github.com/danielmiessler/Fabric/blob/b682dad740f24e85ce9a48d23babc6780dd476ac/data/patterns/extract_recommendations/system.md", "repository": "https://github.com/danielmiessler/Fabric", "revision": "b682dad740f24e85ce9a48d23babc6780dd476ac", "path": "data/patterns/extract_recommendations/system.md", "license": "MIT", "license_url": "https://github.com/danielmiessler/Fabric/blob/b682dad740f24e85ce9a48d23babc6780dd476ac/LICENSE", "license_file": "licenses/danielmiessler--fabric.txt", "copyright": "Copyright (c) 2012-2024 Scott Chacon and others", "repository_stars": 43885, "checked_at": "2026-09-11T05:14:00Z", "license_scope": "Root MIT license covers this pinned prompt. No nested license or NOTICE applies.", "adaptation": "Retains up to 20 recommendations at 16 words each and permits source-grounded inference. Adds explicit inference labeling and conditions instead of generic advice." } ], "instructions": "Extract the practical recommendations in the supplied content. Include explicit recommendations and useful actions that follow directly from the source.\n\nReturn only a bulleted list of up to 20 recommendations. Keep each bullet within 16 words when that preserves the necessary meaning. Begin with a clear action verb and remove repeated advice.\n\nRetain the conditions, timing, responsible people, and limits that determine when an action applies. Label an inferred recommendation as \"Inferred\" so it cannot be mistaken for the author's words. Do not add outside advice, new commitments, or unsupported causes. If the source contains no recommendations, say so briefly instead of inventing a list.", "example": { "context": "Service retrospective with release gates", "before": "The retrospective recommends running the current service's backup restore test once each month. It also recommends giving every failed restore test a named owner by the next business day, so failures remain tracked. Before a release that changes stored data, the team should confirm that the latest restore test passed. If that test failed, deployment should stay paused until the failure is resolved. These recommendations apply only to the current service. The retrospective makes no recommendation to change the test schedule for other systems.", "after": "- Run the current service's backup restore test monthly.\n- Assign each failed restore test a named owner by the next business day.\n- Confirm the latest restore test passed before releases that change stored data.\n- Keep deployment paused until a failed restore test is resolved.", "must_preserve": [ "Recommendations scoped to current service", "Monthly restore test", "Named owner for every failed restore test by next business day", "Data-changing releases require latest restore test to have passed", "Failed latest test keeps deployment paused until resolved", "All recommendations explicit, not inferred" ], "omitted": [ "Drops rationale and a repeated scope statement. No recommendation for other services is inferred." ], "kind": "illustrative" } }