research-and-ingest

This skill is invoked by research-scout to fetch external content in a disciplined way. The output of this skill is two artifacts: raw snapshots in docs/graph/sources/raw/ (when license allows) and normalized summaries in docs/graph/sources/normalized/, plus a row in docs/graph/sources/index.md. The wiki page (created by library-wiki) draws from these.

When to apply this skill

• A new dependency is being evaluated or added.
• A wiki page's "Last reviewed" date is older than the project's review cadence, or its pin no longer matches the lockfile.
• An ADR is being written and needs current evidence.
• An LLM/VLM feature is being designed and needs the provider's current behavior documentation.
• A spec is being authored and refers to a standard (RFC, schema, protocol) the project hasn't read recently.

Source ranking

When two sources disagree, prefer in this order:

1. Official upstream documentation for the exact version in use.
2. Official upstream source code (especially public API surface, examples directory, CHANGELOG).
3. Official upstream blog posts and migration guides.
4. Security advisories from trusted bodies (CVE, CISA, OWASP, official upstream advisories).
5. Well-maintained community resources with current dates.
6. Recent blog posts from credible authors.
7. Anything else, marked clearly with reliability community or mirror.

Never cite a forum answer older than a year for a fast-moving library without verifying against current docs. Never paste a forum answer into the wiki without testing it.

Workflow (per source)

1. Identify

For each source you intend to ingest, note:

• Authority (who maintains it).
• Version coverage (which versions of the library/spec it covers).
• Date (when it was last updated upstream).
• License (whether snapshotting is allowed).
• Slug (the filename you'll use locally).

2. Fetch

Use the host tool's web-fetch capability. If a documentation MCP server is configured (Context7, DeepWiki, llms.txt provider, similar), prefer it for fast, version-aware retrieval. The MCP server does not replace the wiki; it just gets you upstream content faster than crawling.

3. Snapshot (when allowed)

When the license permits, write the raw content to docs/graph/sources/raw/<slug>-<retrieved-date>.<ext>. Acceptable extensions: .md, .html, .pdf, .txt, .json. Strip nothing from the raw file — preserve provenance. When the license does not permit — or the host retrieved through an MCP summary and holds no page to keep — say so in the raw: line of the normalized metadata block. A snapshot with neither the raw file nor the reason is UNJUSTIFIED at the coverage gate: the reason recorded is what makes the omission a decision instead of a habit.

4. Normalize

Produce docs/graph/sources/normalized/<slug>.md:

• Clean Markdown, no navigation chrome, no ads, no tracking pixels, no boilerplate footers.
• Keep upstream headings; drop everything else that is not decision-relevant.
• At the top, the metadata block:

---
source-title: <title>
source-url: <url>
source-maintainer: <name>
source-version-coverage: <e.g. "1.4 - 1.6">
retrieved: YYYY-MM-DD
license: <SPDX or "see source">
raw: raw/<slug>-<retrieved-date>.<ext> | withheld — <why: the license, a host without fetch, an MCP summary with no page behind it>
reliability: official | community-trusted | community | mirror
---

5. Register

Add a row to docs/graph/sources/index.md:

| <title> | <url> | <maintainer> | YYYY-MM-DD | <version> | <reliability> | <relevance> | <one-line notes> |

6. Draft, then hand back

Draft the wiki page from the normalized sources (skill.library-wiki says what each section owes) and end the turn with the handback payload naming the page, the sources index rows, and tester as recommended_next for the smoke test — the scout is a leaf; the opus-class docs-librarian finalizes the page in the close-out, the phase order being ingest-library.flow's.

Documentation MCP servers (when available)

If the project has any of these configured, prefer them for fetching upstream content:

• Context7 (@upstash/context7-mcp): current docs for many libraries, addressable by library ID and version. Use the library-ID form for precision.
• DeepWiki: open-source repository summaries.
• llms.txt providers: projects that publish a machine-readable docs index.

When you use one, cite the source in the normalized file's metadata with the MCP server name and the date.

The local wiki is still authoritative for the project. The MCP server gets you upstream content faster; the wiki page is your distillation of what this project actually does with that content.

Disagreement handling

If two sources disagree:

1. Note the version coverage of each.
2. Prefer the more recent official source.
3. If a security advisory disagrees with the docs, the advisory wins.
4. If the disagreement persists, record both with their versions in the wiki page, and open a question in grill.md §12.

Source reconciliation (lightweight drift check)

Between full research passes, run a cheap periodic reconciliation: diff the currently-resolved dependency versions and manifests/locks against the versions recorded in the library wiki, without re-running research or re-fetching upstream. Classify each line:

• no mismatch — the wiki pin still matches what resolves; nothing to do.
• refresh before the next API-affecting change — the pin has drifted but no work is about to touch that surface; flag it, don't re-ingest yet.
• superseded — treat as historical — the recorded version is gone from the resolved set; mark the wiki content as historical.

This catches silent pin drift that accumulates between full passes, at a fraction of the cost. It is distinct from a full research pass (which re-fetches and re-normalizes upstream content) and from validate-knowledge (which tests whether the wiki prose is navigable and correct, not whether its pins are still current).

Anti-patterns

• Ingesting paywalled or login-walled content without the user's explicit OK.
• Paraphrasing past the point where the paraphrase still says what the source said.
• Skipping the version pin. "Latest" is not a version.
• Snapshotting forbidden content. When the license disallows snapshotting, link only; do not store.
• Ingesting from one source per topic. For non-trivial topics, cross-check against the source code.

Reference files

• docs/graph/templates/library-page.template.md — where this skill's output ultimately lands.
• docs/graph/protocols/ingest-library.md — the parent protocol.
• docs/graph/agents/10-research-scout.md — the agent that runs this.
• docs/graph/skills/library-wiki.md — the wiki-maintenance skill.