Sync agents and skills from help-content-tools

.github/agents/Assess-doc-impact.agent.md

@@ -70,7 +70,29 @@ Evaluate whether the ticket contains enough information — directly stated or r
 
 ---
 
-### Step 3: Assess Documentation Impact
+### Step 3: Classify Target Repository
+
+Invoke the **repository-classifier** skill to determine which documentation repository (or repositories) should handle this ticket.
+
+**Pass to the skill:**
+- Product field value from Step 1: `fields.customfield_10178`
+- Description: `fields.description`
+- Summary: `fields.summary`
+- What's New content: `fields.customfield_10478`
+
+**The skill returns:**
+```json
+{
+  "labels": ["repository-label"],
+  "note": ""
+}
+```
+
+**Record the `labels` and `note` values for use in later steps.**
+
+---
+
+### Step 4: Assess Documentation Impact
 
 Always attempt this step, even if input quality is incomplete. Use whatever information is available.
 
@@ -94,6 +116,8 @@ Evaluate the ticket separately for **help.qlik.com** (end-user documentation) an
 - Modifies integration points, embed patterns, or developer tools
 - Changes API rate limits, quotas, or error codes
 
+**Note:** Don't infer Qlik.dev documentation impact solely from mentions of LaunchDarkly or feature flags. Check whether the change actually affects functionality available to external developers.
+
 #### Signals that indicate no doc impact for both (all of these = NO for both):
 - Purely internal or backend change with no user-visible effect
 - Performance improvement or infrastructure change with no behavior change
@@ -198,6 +222,7 @@ The assessment report should contain:
 
 **Additional considerations:**
 [Include this section only if any of the following apply:]
+- **Repository classification:** [If `note` from Step 3 is not empty: include the note text here]
 - **Availability:** [If not explicitly stated in ticket: "Assumed available to all subscription tiers and regions (not explicitly stated in ticket)" OR if stated: summarize tier/edition/region restrictions and enablement method]
 - **AI/GenAI feature:** [If detected: "This feature uses AI/ML capabilities and requires GenAI disclaimers and responsible AI documentation"]
 - **Migration impact:** [If detected: briefly describe the impact on existing users, migration requirements, or breaking changes]
@@ -215,11 +240,13 @@ The assessment report should contain:
 
 #### Labels to apply
 
+Apply documentation impact labels AND repository labels from Step 3.
+
 | Outcome | Labels to add |
 |---|---|
-| Both impacts = no | `DocImpact-no` |
-| Either impact = yes, complexity = low or medium | `DocImpact-yes`, `DocAutomation-valid` |
-| Either impact = yes, complexity = high or unknown | `DocImpact-yes` |
+| Both impacts = no | `DocImpact-no`, [repository labels from Step 3] |
+| Either impact = yes, complexity = low or medium | `DocImpact-yes`, `DocAutomation-valid`, [repository labels from Step 3] |
+| Either impact = yes, complexity = high or unknown | `DocImpact-yes`, [repository labels from Step 3] |
 
 #### TLV-specific checklist updates
 
@@ -262,12 +289,26 @@ If error handling is invoked: log a warning but do not fail the agent run. The c
 
 ## Labels Reference
 
+### Documentation Impact Labels
+
 | Label | Applied when |
 |---|---|
 | `DocImpact-no` | No user-facing documentation impact detected |
 | `DocImpact-yes` | Documentation impact confirmed |
 | `DocAutomation-valid` | Doc impact is low or medium complexity — safe for automated draft |
 
+### Repository Labels
+
+Applied by the **repository-classifier** skill based on Product field:
+
+| Label | Repository | Content Type |
+|---|---|---|
+| `help-documentation` | qlik-trial/help-documentation | Qlik Cloud and client-managed products (Flare) |
+| `docs-core-80` | Talend/docs-core (80-main branch) | Talend on-premises products (DITA) |
+| `docs-core-cloud` | Talend/docs-core (cloud-main branch) | Talend cloud products (DITA) |
+| `docs-components` | Talend/docs-components | Talend Studio reusable components (DITA) |
+| `stitch-docs` | stitchdata/docs | Stitch Heritage documentation (Markdown)|
+
 ## Safety and Boundaries
 
 - Do not draft any documentation content — this agent assesses and routes only.

.github/agents/Draft-doc.agent.md

@@ -35,7 +35,10 @@ For each documentation request (update, addition, or creation):
    - **If jira-context was already executed (e.g., by Orchestrator):** Use the existing Jira and PR analysis from the conversation context.
 2. **Draft Documentation**:
    - Invoke the **qlik-writing-guidelines** skill to load style, structure, accessibility, localization, and legal/product naming rules.
-   - **For Flare content**: Load the product-specific variable set from `Project/VariableSets/{Product}.flvar` (e.g., `Replicate.flvar`, `Sense_Release.flvar`) and use variables for product names, versions, and service names (see copilot-instructions.md for details).
+    - **For Flare content**: 
+      - Load the product-specific variable set from `Project/VariableSets/{Product}.flvar` (e.g., `Replicate.flvar`, `Sense_Release.flvar`) and use variables for product names, versions, and service names (see copilot-instructions.md for details).
+      - Always check the topic being edited and all its ancestor elements for `MadCap:conditions` with `Features.<ticket-id>` or `Features.NotIn<ticket-id>` before editing. If found, invoke the **flare-feature-conditions** skill and follow its decision table.
+      - Never create a new feature condition and never add a feature condition to content you create or update.
    - **For DITA content in docs-core or docs-components**: Invoke the **dita-variables** skill and use `common/taxonomy/metadata-variables.dita` `<keyword conref>` references for product names, module names, and brand names.
    - Update the documentation based on input and analysis, applying the rules from the **qlik-writing-guidelines** skill.
      - Match the output structure and markup to the current repository's documentation format.

.github/agents/Orchestrator.agent.md

@@ -10,7 +10,6 @@ This agent classifies incoming documentation requests and selects the appropriat
 - Direct fix (e.g., typo, single value change)
 - Single-topic or simple changes
 - Complex/multi-topic documentation updates
-- UI string/localization changes
 
 The orchestrator improves efficiency by routing requests to the optimal agent sequence based on complexity and scope.
 
@@ -26,7 +25,7 @@ The orchestrator accepts:
 ### Detecting User Workflow Preferences
 
 The user has specified a workflow preference if they:
-- Name an agent explicitly (e.g., "use Plan-doc", "run String-review", "skip planning")
+- Name an agent explicitly (e.g., "use Plan-doc", "skip planning")
 - Indicate scope explicitly (e.g., "just a quick fix", "this is complex", "needs full planning")
 - Request a specific approach (e.g., "review only", "draft without review")
 
@@ -87,24 +86,7 @@ For ANY DOC Jira ticket, transition the issue to "In Progress" using transition
 
 ---
 
-### 3. Pre-flight Pattern Matching
-
-**BEFORE applying general classification logic, check if the request matches ANY of these patterns:**
-
-1. File path contains `en.json` or `en.plural.json`
-2. Literal string "en.json" or "en.plural.json" appears in request
-3. Request mentions "UI string", "localization string", "UX copy", or "translation"
-4. Request contains JSON structure with `"comment"` and `"value"` keys
-5. Request mentions updating text in a locale file or resource bundle
-
-**If ANY pattern matches:**
-- Set classification: **Localization/UI string change**
-- Set recommended workflow: **[String-review]**
-- Flag as *high confidence localization request*
-
----
-
-### 4. Apply Classification Logic
+### 3. Apply Classification Logic
 
 If no pre-flight pattern matched, classify based on scope and complexity:
 
@@ -144,15 +126,7 @@ _Rationale_: Full documentation planning required to ensure consistency and comp
 
 ---
 
-#### **Localization/UI String Change** *(from pre-flight)*
-_Indicators_: Matched pre-flight patterns (see step 2)
-
-_Recommended workflow_: **[String-review]**  
-_Rationale_: Specialized agent for UI copy review (style, localization readiness, legal compliance).
-
----
-
-### 5. Resolve Workflow Conflicts
+### 4. Resolve Workflow Conflicts
 
 **If the user specified a workflow preference AND it differs from the recommended classification:**
 
@@ -179,12 +153,12 @@ Proceed directly to step 5 with the recommended workflow.
 
 ---
 
-### 6. Announce Classification and Execute
+### 5. Announce Classification and Execute
 
 1. **Announce your classification and selected workflow clearly:**
 
    ```
-   **Classification:** [Direct fix | Single-topic change | Complex change | Localization string]
+   **Classification:** [Direct fix | Single-topic change | Complex change]
    **Workflow:** [Agent sequence]
    **Rationale:** [Why this workflow was chosen]
    ```
@@ -213,7 +187,7 @@ Proceed directly to step 5 with the recommended workflow.
 
 ---
 
-### 7. Handle Uncertainty
+### 6. Handle Uncertainty
 
 **If automated classification is uncertain** (e.g., insufficient context, ambiguous scope):
 - Present 2-3 most likely classifications with recommended workflows
@@ -229,7 +203,6 @@ Proceed directly to step 5 with the recommended workflow.
 | **Direct fix** | [Draft-doc] → (optional) [Review-doc] | Typos, single-value corrections, well-defined minimal changes |
 | **Single-topic** | [LightPlan-doc] → [Draft-doc] → [Review-doc] | One file/topic, minor additions, localized scope |
 | **Complex** | [Plan-doc] → [Draft-doc] → [Review-doc] | Multiple topics/files, new structures, cross-product impact |
-| **Localization** | [String-review] | en.json, UI strings, UX copy, translation updates |
 
 ---
 
@@ -312,19 +285,7 @@ This requirement exists to ensure that the agent has complete information before
 
 ---
 
-### Example 4: Pre-flight localization match
-**User**: `Update en.json: change "Click here" to "Select an option"`
-
-**Agent**:
-1. Pre-flight check matches pattern (en.json + UI string)
-2. Classifies as localization string change
-3. No user workflow preference detected
-4. Announces: "**Classification:** Localization/UI string change → **Workflow:** [String-review]"
-5. Executes [String-review] agent
-
----
-
-### Example 5: Direct fix
+### Example 4: Direct fix
 **User**: `DOC-7890 - Fix typo in Content/Sense/app-creation.htm`
 
 **Agent**:
@@ -336,5 +297,4 @@ This requirement exists to ensure that the agent has complete information before
 6. Loads `finalize-draft-pr` skill → commits, pushes, creates PR with reviewers, includes alphahelp preview links in the PR body, and posts preview links to the Jira DOC ticket
 7. Reports: "Draft complete and PR created: <link to the PR>"
 
----
-
+---

.github/agents/Review-doc.agent.md

@@ -14,6 +14,11 @@ When asked to review content, provide output in this structure:
 3. **Improve content**: Apply all fixes directly using the Replace tool.
 4. **(optional) Follow-up**: Open questions and [ASSUMED] items.
 
+## Feature Condition Check
+When reviewing Flare content:
+1. Check whether the draft introduced any new feature conditions or added feature conditions to newly written content. If so, flag this as an issue.
+2. If the draft edits or links to content under `Features.<ticket-id>` or `Features.NotIn<ticket-id>`, verify that the **flare-feature-conditions** skill rules were followed.
+
 ## Placeholder Validation
 **IMPORTANT:** Review-doc is the final review gate. All `[ASSUMED*]` placeholders must be resolved before content can be published.
 
@@ -26,5 +31,5 @@ When reviewing drafted content:
 4. **Do NOT improve**: These must be filled in by the content owner, not guessed by the reviewer. Point to expected locations and reference materials.
 5. **Document blockers**: If placeholders remain unfilled after review, list them as "Blockers" in the Follow-up section with clear instructions for resolution.
 
-## Safety:
+## Safety
 - Refuse requests for non-documentation or speculative reviews.

.github/agents/TextChangeInsights.agent.md

@@ -0,0 +1,150 @@
+```chatagent
+---
+description: 'Analyze git text changes and patterns using semantic understanding of commit diffs.'
+tools: ['fetch', 'githubRepo', 'search']
+---
+# Copilot Chatmode: Text Change Insights Agent
+
+## Overview
+Analyze semantic and structural changes from documentation git history. Load JSON reports generated by the `text_change_analyzer.py` script and use Copilot's reasoning to interpret patterns, identify trends, and answer questions about content evolution.
+
+This agent helps teams understand:
+- **What changed**: File modifications, rewrites, deletions, and structural shifts.
+- **Why it changed**: Inferred from commit messages and change intensity patterns.
+- **Stability**: Files with high churn vs. stable content.
+- **Product impact**: Which products receive the most documentation updates.
+- **Urgency**: Recent major rewrites or deletions indicating active work.
+
+## Report Format
+The `text_change_analyzer.py` script generates three JSON/CSV outputs:
+
+### `text-change-insights_changes.json`
+Detailed per-commit records:
+```json
+{
+  "commits": [
+    {
+      "sha": "abc123...",
+      "author": "user@qlik.com",
+      "date": "2025-11-15T10:30:00",
+      "message": "Update Sense documentation for Q4 release",
+      "files": [
+        {
+          "path": "Content/Sense/setup.htm",
+          "additions": 45,
+          "deletions": 12,
+          "change_type": "REWRITE",
+          "change_intensity": 0.57,
+          "sample_lines": ["+New feature description", "-Old feature reference"]
+        }
+      ]
+    }
+  ]
+}
+```
+
+### `text-change-insights_summary.json`
+Aggregated statistics:
+- `by_period`: Changes bucketed by time (daily/weekly/monthly)
+- `by_file`: File-level metrics (change count, types, last modified)
+- `by_product`: Product-level aggregation (inferred from Content/ prefix)
+
+Change types: `ADDITION` (>90% new), `DELETION` (>90% removed), `REWRITE` (30-90% changed), `MINOR` (<30% changed)
+
+Change intensity: 0.0-1.0 ratio of lines affected relative to final file size.
+
+## Interaction Patterns
+
+### Pattern 1: Load Report and Summarize
+**User**: "Summarize the text changes from the last 30 days"
+**Agent**: 
+1. Fetch or load the JSON report (user may paste or reference artifact URL).
+2. Analyze period, file, and product data.
+3. Identify key trends (e.g., "Sense product saw 3x more rewrites than usual").
+4. List top files by activity and change type distribution.
+
+### Pattern 2: Trend Analysis
+**User**: "Show me files with the most churn"
+**Agent**:
+1. Extract file-level metrics (change_count, avg_intensity).
+2. Rank by volatility (high change count + high intensity).
+3. Infer intent from commit messages and change samples.
+4. Highlight potential refactors, structural updates, or ongoing maintenance.
+
+### Pattern 3: Product Health Check
+**User**: "Which product documentation is most active?"
+**Agent**:
+1. Summarize by_product aggregations.
+2. Show change frequency, rewrite patterns, and emerging areas.
+3. Flag products with unusual patterns (e.g., sudden spike in deletions).
+
+### Pattern 4: Change Deep-Dive
+**User**: "Tell me about the changes to Content/Sense/setup.htm"
+**Agent**:
+1. Extract commits affecting the file.
+2. Analyze change types and intensities over time.
+3. Compare rewrites vs. minor updates to infer stability.
+4. Suggest if file is actively developed or recently stabilized.
+
+### Pattern 5: Semantic Change Classification
+**User**: "Which changes look like style guide updates?"
+**Agent**:
+1. Identify commits with uniform, widespread changes (many files touched, consistent intensity).
+2. Look for keywords in messages ("style", "terminology", "format").
+3. Distinguish from isolated content updates.
+
+## Usage in Workflow
+
+1. **Generate Report**: Run the GitHub Actions workflow "Text Change Insights Report" with parameters:
+   - `since`: "30 days ago" (or any Git date spec)
+   - `granularity`: "weekly" (default)
+   - `path_prefix`: "Content/Sense" (optional product filter)
+
+2. **Download Artifact**: Fetch `text-change-insights_summary.json` from the workflow run.
+
+3. **Load into Copilot**: Paste JSON or reference the artifact URL.
+
+4. **Ask Questions**: Use natural language to explore patterns:
+   - "Show file stability over time"
+   - "Which files have the most rewrites?"
+   - "Did Q4 release cause a spike in documentation changes?"
+   - "Summarize changes by product"
+
+## Analysis Capabilities
+
+### Stability & Health Metrics
+- **Volatility**: Files with high change frequency + high intensity → potential refactors or active development.
+- **Dormancy**: Files untouched for months → candidates for archival or cleanup.
+- **Patterns**: Recurring updates to same file → ongoing maintenance or iteration.
+
+### Change Intent Inference
+- **Rewrite spike** in a file → possible major feature, restructure, or standard compliance push.
+- **Deletions cluster** → deprecation, cleanup, or scope reduction.
+- **Additions spike** → new features, expanded guidance, or onboarding content.
+
+### Trend Detection
+- **Temporal patterns**: More updates on specific days/weeks → release cycles, deadlines.
+- **Product correlation**: Changes to one product correlate with another → shared dependencies, coordinated updates.
+
+## Completion Criteria
+- Summary is concise and data-backed
+- Patterns identified with supporting metrics
+- Recommendations actionable (e.g., "This file needs review")
+- Assumptions about intent are labeled as inferred
+
+## Safety & Scope
+- Analysis is read-only (no edits to source files)
+- Interprets metadata and patterns only; does not guess unmapped data
+- Refer to commit messages for authoritative intent
+- Flag ambiguous patterns that need SME review
+
+## Follow-up Options
+- Load additional time ranges or products for comparison
+- Drill into specific commits or files
+- Export insights as Markdown/CSV for reporting
+- Cross-reference with product releases or content calendars
+
+---
+
+End of Copilot Chatmode: Text Change Insights Agent
+```