@grainulation/wheat 1.0.2 → 1.0.4

package/templates/commands/sync.md

@@ -0,0 +1,94 @@
+# /sync — Publish sprint artifacts to external platforms
+
+You are publishing the current Wheat sprint's artifacts to external platforms. Read CLAUDE.md for sprint context and claims.json for the current state.
+
+## Process
+
+1. **Compile first**: Run `wheat compile --summary` to ensure compilation is fresh. If compilation is blocked, stop and tell the user to fix issues first.
+
+2. **Detect available targets**: Check which MCP servers are available:
+
+   - **Confluence** (via Atlassian MCP): Push the compiled brief as a Confluence page. Use `mcp__atlassian__confluence_create_page` or `mcp__atlassian__confluence_update_page`.
+   - **Slack** (via Slack MCP): Post a sprint summary to a channel. Use `mcp__slack__send_message`.
+   - **Notion** (via Notion MCP): Create a page with the brief content.
+   - **Local file export**: Always available — write to `output/` directory.
+
+3. **Confluence publish** (primary target):
+
+   - Read `compilation.json` for the certified output
+   - Read `output/brief.html` if it exists, otherwise generate the brief first via `/brief`
+   - Convert the brief to Confluence Storage Format (XHTML subset):
+     - Strip `<script>` and `<style>` tags
+     - Convert inline styles to Confluence macros where possible
+     - Wrap code blocks in `<ac:structured-macro ac:name="code">`
+     - Preserve tables, headings, lists as-is (Confluence supports standard HTML)
+   - Create or update the page:
+     - Page title: `[Sprint] <question summary>` (truncated to 255 chars)
+     - Space key: from user argument or CLAUDE.md connector config
+     - Add labels: `wheat-sprint`, `<phase>`, claim type tags
+   - Add a comment with sync metadata: timestamp, claims hash from compilation certificate, claim count
+
+4. **Build sync manifest**: After publishing, append a sync record to `output/sync-log.json`:
+
+   ```json
+   {
+     "timestamp": "<ISO>",
+     "target": "confluence|slack|notion|file",
+     "url": "<published URL or file path>",
+     "claims_hash": "<from compilation.json>",
+     "claims_count": <number>,
+     "phase": "<current phase>"
+   }
+   ```
+
+5. **Add a sync claim**: Append a claim recording the sync event:
+
+   ```json
+   {
+     "id": "r<next>",
+     "type": "factual",
+     "topic": "sync",
+     "content": "Sprint artifacts published to <target> at <url>. Claims hash: <hash>. <count> active claims at time of sync.",
+     "source": {
+       "origin": "mcp",
+       "artifact": "<url>",
+       "connector": "<target>"
+     },
+     "evidence": "documented",
+     "status": "active",
+     "phase_added": "<phase>",
+     "timestamp": "<ISO>",
+     "conflicts_with": [],
+     "resolved_by": null,
+     "tags": ["sync", "<target>"]
+   }
+   ```
+
+6. **Compile again** after adding the sync claim.
+
+## Arguments
+
+- No argument: auto-detect available targets, prefer Confluence
+- `confluence <space-key>`: publish to specific Confluence space
+- `slack <channel>`: post summary to Slack channel
+- `notion <page-id>`: publish to Notion page
+- `file`: export to `output/` only (always works, no MCP needed)
+- `--dry-run`: show what would be published without actually pushing
+
+## Error handling
+
+- If no MCP servers are configured, fall back to local file export and tell the user how to configure Atlassian MCP: `claude mcp add atlassian -- npx @anthropic-ai/atlassian-mcp`
+- If Confluence API returns 403, suggest checking OAuth scopes
+- If the page already exists, update it (don't create duplicates) — search by title first
+
+## Next steps suggestions
+
+After sync completes, suggest:
+
+```
+Next steps:
+  /status              -- verify sync claim was recorded
+  /sync slack <channel> -- cross-post to Slack
+  /pull confluence      -- backfill claims from Confluence comments
+  /brief               -- regenerate brief if claims changed
+```

package/templates/commands/witness.md

@@ -5,6 +5,7 @@ You are corroborating (or contradicting) a specific claim using an external sour
 ## Process
 
 1. **Parse arguments**: The user provides a claim ID and an external URL.
+
    - Example: `/witness p001 https://nodejs.org/api/http.html`
    - If only a claim ID is given, ask for the URL
    - If only a URL is given, ask which claim to witness
@@ -14,13 +15,15 @@ You are corroborating (or contradicting) a specific claim using an external sour
 3. **Fetch the external source**: Use web fetch to read the URL. If it's documentation, source code, or an article, extract the relevant content.
 
 4. **Classify the relationship** between the external evidence and the claim:
+
    - **Full support** -> external source confirms the claim completely
    - **Partial support** -> confirms some assertions but adds caveats
    - **Partial contradiction** -> external source challenges some assertions
    - **Full contradiction** -> external source directly contradicts the claim
 
 5. **Determine evidence tier** based on source type:
-   - Official docs (*.nodejs.org, docs.*, RFC) -> `documented`
+
+   - Official docs (_.nodejs.org, docs._, RFC) -> `documented`
    - Blog posts, Stack Overflow, tutorials -> `web`
    - GitHub source code, changelogs -> `documented`
    - Production metrics, dashboards -> `production`
@@ -52,6 +55,7 @@ You are corroborating (or contradicting) a specific claim using an external sour
 ```
 
 **Important relationship -> action mapping:**
+
 - Full/partial support: No `conflicts_with`. The witness corroborates.
 - Partial contradiction: Set `conflicts_with: ["<target claim ID>"]`. Explain the contradiction clearly in content.
 - Full contradiction: Set `conflicts_with: ["<target claim ID>"]`. This becomes a challenge claim effectively.
@@ -76,4 +80,4 @@ Commit: `wheat: /witness <target claim ID> — added <witness claim IDs> (<relat
 - The corroboration count for the witnessed claim
 - Suggest: `/witness` for other uncorroborated claims, `/challenge` if partial contradiction warrants deeper investigation
 
-$ARGUMENTS
+$ARGUMENTS