Add Wave general-purpose pipelines

ADR, changelog, code-review, debug, doc-sync, explain, feature,
hotfix, improve, onboard, plan, prototype, refactor, security-scan,
smoke-test, speckit-flow, supervise, test-gen, and more.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

.wave/pipelines/ingest.yaml

@@ -0,0 +1,173 @@
+kind: WavePipeline
+metadata:
+  name: ingest
+  description: "Ingest a web article into the Zettelkasten as bibliographic and permanent notes"
+  release: true
+
+input:
+  source: cli
+  examples:
+    - "https://simonwillison.net/2026/Feb/7/software-factory/"
+    - "https://langfuse.com/blog/2025-03-observability"
+    - "https://arxiv.org/abs/2401.12345"
+
+steps:
+  - id: fetch
+    persona: scout
+    workspace:
+      mount:
+        - source: ./
+          target: /project
+          mode: readonly
+    exec:
+      type: prompt
+      source: |
+        Fetch and extract structured content from a web article.
+
+        URL: {{ input }}
+
+        ## Steps
+
+        1. Use WebFetch to retrieve the article content
+        2. Extract:
+           - title: article title
+           - author: author name (look in byline, meta tags, about section)
+           - date: publication date
+           - summary: 50-3000 character summary of the article
+           - key_concepts: list of key concepts with name and description
+           - notable_quotes: direct quotes with context
+           - author_year_key: generate AuthorYear key (e.g., Willison2026)
+        3. If the author name is unclear, use the domain name as author
+
+        ## Output
+
+        Write the result as JSON to output/source-extract.json matching the contract schema.
+    output_artifacts:
+      - name: source-extract
+        path: output/source-extract.json
+        type: json
+    handover:
+      contract:
+        type: json_schema
+        source: output/source-extract.json
+        schema_path: .wave/contracts/source-extract.schema.json
+        on_failure: retry
+        max_retries: 2
+
+  - id: connect
+    persona: navigator
+    dependencies: [fetch]
+    memory:
+      inject_artifacts:
+        - step: fetch
+          artifact: source-extract
+          as: source
+    workspace:
+      mount:
+        - source: ./
+          target: /project
+          mode: readonly
+    exec:
+      type: prompt
+      source: |
+        Find connections between extracted source content and existing Zettelkasten notes.
+
+        Read the source extract: cat artifacts/source
+
+        ## Steps
+
+        1. For each key concept in the source, search for related notes:
+           - `notesium lines --filter="concept_name"`
+           - Read the most relevant matches
+        2. Identify the Folgezettel neighborhood where new notes belong:
+           - What section does this content fit in?
+           - What would be the parent note?
+           - What Folgezettel address should new notes get?
+        3. Check if the index note needs updating
+        4. Determine link directions (should new note link to existing, or existing link to new?)
+
+        ## Output
+
+        Write the result as JSON to output/connections.json matching the contract schema.
+        Include:
+        - source_title: title of the source being connected
+        - related_notes: list of related existing notes with filename, title,
+          folgezettel_address, relationship explanation, and link_direction
+        - suggested_placements: where new notes should go in the Folgezettel
+          with address, parent_note, section, rationale, and concept
+        - index_update_needed: boolean
+        - suggested_index_entries: new entries if needed
+        - timestamp: current ISO 8601 timestamp
+    output_artifacts:
+      - name: connections
+        path: output/connections.json
+        type: json
+    handover:
+      contract:
+        type: json_schema
+        source: output/connections.json
+        schema_path: .wave/contracts/connections.schema.json
+        on_failure: retry
+        max_retries: 2
+
+  - id: create
+    persona: scribe
+    dependencies: [connect]
+    memory:
+      inject_artifacts:
+        - step: fetch
+          artifact: source-extract
+          as: source
+        - step: connect
+          artifact: connections
+          as: connections
+    workspace:
+      mount:
+        - source: ./
+          target: /project
+          mode: readwrite
+    exec:
+      type: prompt
+      source: |
+        Create Zettelkasten notes from an ingested web source.
+
+        Read the artifacts:
+          cat artifacts/source
+          cat artifacts/connections
+
+        ## Steps
+
+        1. **Create the bibliographic note**:
+           - Use `notesium new` for the filename
+           - Title: `# AuthorYear` using the author_year_key from the source extract
+           - Content: source URL, author, date, summary, key quotes
+           - One sentence per line
+
+        2. **Create permanent notes** for key ideas that warrant standalone Zettel:
+           - Use `notesium new` for each
+           - Use the Folgezettel address from suggested_placements
+           - Title: `# {address} {Concept-Name}`
+           - Write in own words — transform, don't copy
+           - Add contextual links to related notes (explain *why* the connection exists)
+           - Link back to the bibliographic note
+
+        3. **Update existing notes** if bidirectional links are suggested:
+           - Add links from existing notes to the new permanent notes
+           - Include contextual explanation for each link
+
+        4. **Update the index note** if index_update_needed is true:
+           - Add new keyword → entry point mappings
+
+        5. **Commit all changes**:
+           - `git add *.md`
+           - `git commit -m "ingest: {AuthorYear key in lowercase}"`
+
+        6. **Write summary** to output/ingest-summary.md:
+           - Bibliographic note created (filename, title)
+           - Permanent notes created (filename, title, Folgezettel address)
+           - Links added to existing notes
+           - Index updates made
+    output_artifacts:
+      - name: ingest-summary
+        path: output/ingest-summary.md
+        type: markdown