content-processing

# Content Processing Pipeline

Canonical pipeline for creating notes from enriched content. This skill is the single source of truth for:

1. **Note creation** — generic `para_create` pattern for all templates
2. **Layer 1 injection** — `para_replace_section` with formatting rules
3. **Commit** — `para_commit` after creation
4. **Null-safety** — field handling rules

Callers (triage-worker, quick-create) handle enrichment, classification, user confirmation, and persistence separately. This skill only covers the "create → inject → commit" pipeline.

---

## Pipeline Overview

```
Enriched content + classification (from caller)
    ↓
1. Discover template metadata (para_template_fields)
    ↓
2. Create note (para_create)
   - Resources: pass Layer 1 via `content` parameter → single call creates + injects + commits
   - Meetings: pass structured body via `content` parameter → single call creates + injects + commits
   - Others: frontmatter-only → auto-commits
```

**Key optimization:** Resources and meetings both use the `content` parameter on `para_create` to inject body content in a single CLI invocation. The CLI internally calls `createFromTemplate → replaceSections → autoCommitChanges` — no separate `para_commit` or `para_replace_section` needed.

---

## Null-Safety Rules

**CRITICAL:** Never pass `null` values in `args` to `para_create`. Omit the key entirely.

```
// WRONG — creates "[[null]]" in frontmatter
args: { projects: null, author: null }

// CORRECT — omit keys with no value
args: { }  // projects and author omitted
```

Apply this rule to ALL optional fields: `projects`, `project`, `author`, `transcription`, `source`.

---

## Enrichment

Before creating notes, content must be enriched. See the canonical routing table and tool selection:

@plugins/para-obsidian/skills/triage/references/enrichment-strategies.md

**Quick summary:** YouTube → `get_transcript`, Articles/GitHub → `firecrawl_scrape`, X/Twitter → X-API MCP (`x_get_tweet`), Voice/Attachment → `