{
  "schema_version": 2,
  "kind": "shortening-method",
  "format": "agent-skill",
  "id": "sparse-priming",
  "name": "Sparse priming",
  "category": "Agent context",
  "summary": "Turn a concept-heavy source into a compact list of meaningful cues.",
  "use_cases": [
    "Concept notes",
    "Agent context",
    "Knowledge summaries"
  ],
  "word_count": 122,
  "url": "https://sho.rten.it/methods/sparse-priming/",
  "instructions_url": "https://sho.rten.it/methods/sparse-priming/SKILL.md",
  "skill_url": "https://sho.rten.it/methods/sparse-priming/SKILL.md",
  "json_url": "https://sho.rten.it/methods/sparse-priming/llms.txt",
  "plain_text_url": "https://sho.rten.it/methods/sparse-priming/prompt.txt",
  "license": "MIT",
  "sources_url": "https://sho.rten.it/sources/#sparse-priming",
  "skill_name": "sparse-priming",
  "skill_description": "Turn a concept-heavy source into a compact list of meaningful cues. Use for Concept notes, Agent context, Knowledge summaries.",
  "agents_md_url": "https://sho.rten.it/methods/sparse-priming/AGENTS.md",
  "sources": [
    {
      "title": "daveshap/SparsePrimingRepresentations · Sparse priming",
      "url": "https://github.com/daveshap/SparsePrimingRepresentations/blob/5d3d19e39bfd222901a5c88d06409b5ef48c1478/system.md",
      "repository": "https://github.com/daveshap/SparsePrimingRepresentations",
      "revision": "5d3d19e39bfd222901a5c88d06409b5ef48c1478",
      "path": "system.md",
      "license": "MIT",
      "license_url": "https://github.com/daveshap/SparsePrimingRepresentations/blob/5d3d19e39bfd222901a5c88d06409b5ef48c1478/LICENSE",
      "license_file": "licenses/daveshap--sparseprimingrepresentations.txt",
      "copyright": "Copyright (c) 2023 David Shapiro",
      "repository_stars": 795,
      "checked_at": "2026-09-10",
      "license_scope": "Source is in the repository or package covered by the linked MIT license; no source-file license override was present.",
      "adaptation": "Retained the source's distilled statements, associations, and concept-cue format. Removed unsupported latent-space and reconstruction claims; no token-efficiency guarantee is made. Added explicit preservation of uncertainty and instructions against reconstructing missing facts. The repository has 795 stars, fewer than other selected source repositories; this is documented rather than inflated."
    }
  ],
  "instructions": "Render the supplied information as a sparse priming representation: a compact list of statements and cues another language model can interpret. Keep the main concepts, relationships, distinctions, and dependencies. Use succinct assertions, associations, and existing analogies when they convey the source efficiently.\n\nPrefer explicit relationships to isolated keywords when the connection would otherwise be ambiguous. Remove repetition and rhetorical framing. Preserve proper names, important numbers, negation, uncertainty, and exceptions. Include enough context to identify what each cue refers to.\n\nDo not invent associations, facts, or metaphors to fill gaps. Do not assume a later model can reconstruct omitted details. If exact wording is required for code or a quotation, preserve it. Return only the cue list, with no explanation of the technique.",
  "example": {
    "before": "The Beacon document index contains manuals from the North and South warehouses, but it does not contain customer records. Each indexed passage stores its source file and page number so that a response can point back to the manual. Manuals are reindexed every night at 02:00 UTC. A manual uploaded after that run will not be searchable until the next successful run. If a run fails, the previous index stays available and an alert is sent to the operations inbox. This means availability of the old index does not establish that the latest manual is present. The team suspects large PDF files contributed to Tuesday's failure, but it has not established the cause. The description repeats that source references help readers check answers.",
    "after": "- Beacon index: North and South warehouse manuals; excludes customer records.\n- Each passage → source file + page number → checkable manual reference.\n- Reindex: nightly, 02:00 UTC.\n- Upload after run → searchable only after next successful run.\n- Failed run → previous index remains available + operations inbox alert.\n- Available index ≠ latest manual included.\n- Tuesday failure: large PDFs suspected; cause unconfirmed.",
    "context": "Fictional retrieval-system context",
    "must_preserve": [
      "Beacon indexes North and South warehouse manuals, not customer records",
      "Each passage retains its source file and page number",
      "Nightly run at 02:00 UTC; later uploads need the next successful run",
      "A failed run retains the previous index and alerts the operations inbox",
      "Availability does not prove freshness",
      "Large PDFs are only a suspected cause of Tuesday’s failure"
    ],
    "omitted": [
      "Repeated explanation of reference value"
    ],
    "kind": "illustrative"
  }
}