From paper-distill
Distills raw paper evidence Markdown files into canonical wiki pages with metadata, key concepts, and wikilinks. Handles single papers or batches using tools like distill_paper and upsert_wiki_page.
npx claudepluginhub hamondyan/paper-distillThis skill uses the workspace's default tool permissions.
Turn captured raw evidence (`raw/evidence/<paper_id>.md`) into a canonical paper page (`wiki/papers/<slug>.md`). This is the cognitive step where an agent reads a long captured document and produces a short, structured, cross-linked summary with required metadata and concept links.
Fetches up-to-date documentation from Context7 for libraries and frameworks like React, Next.js, Prisma. Use for setup questions, API references, and code examples.
Executes ctx7 CLI to fetch up-to-date library documentation, manage AI coding skills (install/search/generate/remove/suggest), and configure Context7 MCP. Useful for current API refs, skill handling, or agent setup.
Generates production-ready Python code for Dataverse SDK with error handling, singleton clients, retry logic, OData optimizations, logging, and type hints.
Share bugs, ideas, or general feedback.
Turn captured raw evidence (raw/evidence/<paper_id>.md) into a canonical paper page (wiki/papers/<slug>.md). This is the cognitive step where an agent reads a long captured document and produces a short, structured, cross-linked summary with required metadata and concept links.
This skill runs after paper-intake has captured the evidence. For read/index work, use docs/qmd-cli.md and runtime qmd --help.
| Tool | Purpose |
|---|---|
distill_paper | Resolve raw evidence and write one canonical paper page from a distilled payload |
distill_papers | Apply the same distillation write workflow sequentially to a batch |
upsert_wiki_page | Write the canonical paper page with write-time schema validation |
check_concept_alias | Resolve concept surfaces before linking |
qmd get, distill_paper(input_value) or by paper_id, read it, extract the structured summary, and call distill_paper(input_value, distilled=...) or upsert_wiki_page(page_type="paper", ...).distill_papers(items) only after each item has grounded distilled content.paper-intake to run ingest_and_read first.Every canonical paper page requires:
paper_id — string, e.g. arxiv:2406.09246.title — full paper title.year — integer publication year.venue — publication venue or arXiv if preprint-only.source_layer — usually wiki.key_concepts_topk — one to five canonical concept surfaces, each wikilinked at least once in the body.The write-time validator rejects pages that violate any of these constraints. Do not attempt to work around the validator by hand-editing files.
qmd get the raw evidence file for the paper.key_concepts_topk using the rules in references/concept-selection.md — prefer reusing existing concepts, call check_concept_alias when uncertain.upsert_wiki_page(page_type="paper", target=<slug>, frontmatter=..., body=...).key_concepts_topk is new to the vault and meets the bar for a concept page, create it via upsert_wiki_page(page_type="concept", ...) in a follow-up call.qmd update after the write round; recommend qmd embed -f only if semantic retrieval must reflect the new page immediately.key_concepts_topk entry must appear as [[Concept Name]] at least once in the body.raw/evidence/ during distillation — it is the preserved source and must stay intact.Relations section so future qmd query runs can surface the comparison.This skill is responsible for reading raw evidence, extracting structure, selecting concepts, and producing pages that satisfy write-time validation.
Python tools are responsible for deterministic writes, schema enforcement, and structured errors.