@thierrynakoa/fire-flow 12.2.1 → 13.0.1

package/commands/fire-test.md

@@ -65,7 +65,7 @@ Step 1: Parse Arguments
 └── Validate command name if --command specified
 
 Step 2: Create Temp Test Directory
-├── Create isolated test directory: $TEMP/fire-flow-test-{timestamp}/
+├── Create isolated test directory: $TEMP/dominion-flow-test-{timestamp}/
 ├── Initialize minimal project structure
 ├── Copy test fixtures from tests/fixtures/
 └── Set up clean environment variables
@@ -181,7 +181,7 @@ Fix: [recommended action]
 **Test Environment:**
 - OS: [Windows/macOS/Linux]
 - Shell: [PowerShell/Bash/Zsh]
-- Plugin Path: ~/.claude/plugins/fire-flow/
+- Plugin Path: ~/.claude/plugins/dominion-flow/
 - Temp Directory: [path]
 ```
 
@@ -429,7 +429,7 @@ tests/fixtures/
 
 | Error | Cause | Fix |
 |-------|-------|-----|
-| "Plugin not found" | fire-flow not installed | Install plugin to ~/.claude/plugins/fire-flow/ |
+| "Plugin not found" | dominion-flow not installed | Install plugin to ~/.claude/plugins/dominion-flow/ |
 | "Skills library empty" | Skills not copied | Copy skills from WARRIOR to skills-library/ |
 | "Hook not firing" | hooks.json misconfigured | Check hooks/hooks.json syntax |
 | "Permission denied" | File permissions | Check write permissions on test directory |
@@ -445,7 +445,7 @@ tests/fixtures/
 ## Running Tests
 
 ### Prerequisites:
-1. Dominion Flow plugin installed at ~/.claude/plugins/fire-flow/
+1. Dominion Flow plugin installed at ~/.claude/plugins/dominion-flow/
 2. Skills library populated (or will use fixtures)
 3. Write access to temp directory
 4. Write access to ~/.claude/warrior-handoffs/
@@ -484,7 +484,7 @@ When `/fire-test` is invoked:
 2. **Set up test environment**:
    ```bash
    # Create temp directory
-   TEST_DIR="$TEMP/fire-flow-test-$(date +%s)"
+   TEST_DIR="$TEMP/dominion-flow-test-$(date +%s)"
    mkdir -p "$TEST_DIR"
 
    # Copy fixtures

package/commands/fire-todos.md

@@ -288,7 +288,7 @@ Return to list.
 After any todo action, update persistent state:
 
 ```markdown
-## .claude/fire-flow.local.md
+## .claude/dominion-flow.local.md
 
 ### Todos
 - Pending: [count]

package/commands/fire-update.md

@@ -56,7 +56,7 @@ The plugin must have a GitHub repository configured in `plugin.json`:
 
 ```json
 {
-  "repository": "https://github.com/username/fire-flow"
+  "repository": "https://github.com/username/dominion-flow"
 }
 ```
 
@@ -66,7 +66,7 @@ If repository is not configured:
 
 To enable updates, add repository URL to plugin.json:
 {
-  "repository": "https://github.com/username/fire-flow"
+  "repository": "https://github.com/username/dominion-flow"
 }
 ```
 
@@ -77,7 +77,7 @@ To enable updates, add repository URL to plugin.json:
 ### Step 1: Validate Environment
 
 ```bash
-PLUGIN_DIR="$HOME/.claude/plugins/fire-flow"
+PLUGIN_DIR="$HOME/.claude/plugins/dominion-flow"
 VERSION_FILE="$PLUGIN_DIR/version.json"
 PLUGIN_JSON="$PLUGIN_DIR/.claude-plugin/plugin.json"
 ```
@@ -300,8 +300,8 @@ git reset --hard "$TARGET_COMMIT"
 ```bash
 # Generate token at: https://github.com/settings/tokens
 # Set token in git remote
-cd ~/.claude/plugins/fire-flow
-git remote set-url origin https://YOUR_TOKEN@github.com/username/fire-flow.git
+cd ~/.claude/plugins/dominion-flow
+git remote set-url origin https://YOUR_TOKEN@github.com/username/dominion-flow.git
 ```
 
 ---

package/commands/fire-validate-skills.md

@@ -0,0 +1,282 @@
+---
+name: fire-validate-skills
+description: Validate skills library structural integrity — frontmatter, sections, cross-references, markdown syntax
+arguments:
+  - name: scope
+    description: "What to validate: 'all', a category name (e.g., 'database-solutions'), or a specific skill filename"
+    required: false
+    type: string
+triggers:
+  - "validate skills"
+  - "check skills"
+  - "skill validation"
+  - "verify skills"
+allowed-tools:
+  - Bash
+  - Read
+  - Write
+  - Edit
+  - Glob
+  - Grep
+  - Agent
+---
+
+# /fire-validate-skills — Skills Library Structural Validator
+
+Validate that skills in the library are well-formed, internally consistent, and have valid cross-references.
+
+## Purpose
+
+Run after `/fire-add-new-skill`, `/fire-research`, or any bulk skill creation to catch:
+- Missing or malformed YAML frontmatter
+- Missing required sections
+- Broken cross-references (related skills pointing to nonexistent files)
+- Unclosed code blocks or malformed tables
+- Invalid field values (bad difficulty, tags as strings instead of arrays)
+
+## Arguments
+
+| Argument | Required | Default | Description |
+|----------|----------|---------|-------------|
+| `scope` | No | `all` | `all` = entire library, `{category}` = one category folder, `{filename.md}` = single file |
+
+## Usage
+
+```bash
+/fire-validate-skills                          # Validate entire library
+/fire-validate-skills database-solutions       # Validate one category
+/fire-validate-skills orm-schema-portability   # Validate single skill
+```
+
+---
+
+## Process
+
+### Step 1: Determine Scope
+
+```
+SKILLS_ROOT = ~/.claude/plugins/dominion-flow/skills-library/
+
+IF scope == "all" or not provided:
+  target_files = glob("{SKILLS_ROOT}/**/*.md") excluding SKILLS-INDEX.md
+ELIF scope matches a directory name in SKILLS_ROOT:
+  target_files = glob("{SKILLS_ROOT}/**/{scope}/*.md")
+ELIF scope matches a filename (with or without .md):
+  target_files = find file by name across all categories
+ELSE:
+  Display: "Unknown scope: {scope}. Use 'all', a category name, or a skill filename."
+  STOP
+```
+
+### Step 2: Validate Each Skill
+
+For each skill file, run these checks:
+
+#### Check 1: YAML Frontmatter (REQUIRED)
+
+```
+Required fields:
+  - name: string (must match filename without .md, kebab-case)
+  - category: string (must match parent directory name)
+  - version: string (semver format: X.Y.Z)
+  - contributed: string (YYYY-MM-DD date format)
+  - contributor: string (non-empty)
+  - last_updated: string (YYYY-MM-DD date format)
+  - tags: array (NOT a string — must be [tag1, tag2, ...])
+  - difficulty: string (must be one of: easy, medium, hard)
+
+Scoring:
+  All present and valid = PASS
+  Missing any required field = FAIL
+  Invalid value (e.g., difficulty: "extreme") = WARN
+```
+
+#### Check 2: Required Sections
+
+```
+Must have ALL of these H2 sections (exact heading text may vary):
+  - Problem / Overview / Purpose (at least one)
+  - Solution Pattern / Complete * / Per-* / How to * (at least one solution section)
+  - When to Use
+  - When NOT to Use
+  - Related Skills (can be empty but must exist)
+  - References (can be empty but must exist)
+
+Scoring:
+  All present = PASS
+  Missing 1-2 = WARN (list which are missing)
+  Missing 3+ = FAIL
+```
+
+#### Check 3: Cross-Reference Resolution
+
+```
+Find all markdown links in "Related Skills" section:
+  Pattern: [text](path.md) or [text](../category/path.md)
+
+For each link:
+  Resolve relative to current file's directory
+  Check if target file EXISTS on disk
+
+Scoring:
+  All resolve = PASS
+  Any broken = FAIL (list broken references with expected path)
+```
+
+#### Check 4: Markdown Syntax
+
+```
+Check for:
+  - Unclosed code blocks (odd number of ``` lines)
+  - Malformed table rows (inconsistent pipe count)
+  - Orphaned heading markers (# at start of line without space after)
+
+Scoring:
+  No issues = PASS
+  Any issue = WARN (list line numbers)
+```
+
+#### Check 5: Content Quality (WARN only, never FAIL)
+
+```
+Check for:
+  - Empty sections (heading with no content before next heading)
+  - Very short skills (< 20 lines total — likely incomplete)
+  - Very long skills (> 500 lines — consider splitting)
+  - Missing code examples in Solution section
+  - Tags array has < 2 tags (too few for searchability)
+
+Scoring:
+  All good = PASS
+  Any issue = WARN (suggestion only)
+```
+
+### Step 3: Generate Report
+
+```markdown
++---------------------------------------------------------------+
+|          SKILLS LIBRARY VALIDATION REPORT                       |
++---------------------------------------------------------------+
+
+Scope: {all | category | single file}
+Files scanned: {N}
+Date: {YYYY-MM-DD}
+
+-------------------------------------------------------------
+RESULTS SUMMARY
+-------------------------------------------------------------
+
+| Status | Count |
+|--------|-------|
+| PASS   | {N}   |
+| WARN   | {N}   |
+| FAIL   | {N}   |
+
+-------------------------------------------------------------
+FAILURES (must fix)
+-------------------------------------------------------------
+
+{For each FAIL:}
+  File: {category}/{filename}.md
+  Check: {which check failed}
+  Issue: {specific problem}
+  Fix: {what to do}
+
+-------------------------------------------------------------
+WARNINGS (should fix)
+-------------------------------------------------------------
+
+{For each WARN:}
+  File: {category}/{filename}.md
+  Check: {which check warned}
+  Issue: {specific problem}
+  Suggestion: {what to improve}
+
+-------------------------------------------------------------
+ALL PASSING ({N} files)
+-------------------------------------------------------------
+
+{List of files that passed all checks — collapsed if > 20}
+
++---------------------------------------------------------------+
+|  VERDICT: {ALL PASS | {N} WARNINGS | {N} FAILURES}            |
++---------------------------------------------------------------+
+```
+
+### Step 4: Route on Verdict
+
+| Verdict | Next Action |
+|---------|------------|
+| ALL PASS | Done — skills are structurally sound |
+| WARNINGS only | Display suggestions, no blocking action needed |
+| FAILURES | List fixes needed. Offer: "Fix these now? [Y/N]" |
+
+If user chooses to fix:
+- Auto-fix simple issues (missing `difficulty: medium` default, missing empty Related Skills section)
+- Flag complex issues for manual review (broken cross-refs need human decision on correct target)
+
+---
+
+## Parallel Execution (for scope=all)
+
+When validating the entire library, spawn parallel agents by category for speed:
+
+```
+Agent per category:
+  - Read all .md files in that category folder
+  - Run Checks 1-5 on each
+  - Return results array
+
+Orchestrator:
+  - Collect results from all agents
+  - Merge into single report
+  - Sort failures first, then warnings, then passes
+```
+
+For small scopes (single category or file), run directly without agents.
+
+---
+
+## Auto-Fix Capabilities
+
+These issues can be auto-fixed with user confirmation:
+
+| Issue | Auto-Fix |
+|-------|----------|
+| Missing `difficulty` field | Add `difficulty: medium` |
+| Missing empty `Related Skills` section | Add `## Related Skills\n\n- None yet\n` |
+| Missing empty `References` section | Add `## References\n\n- None yet\n` |
+| `tags` as string instead of array | Convert `tags: "foo, bar"` → `tags: [foo, bar]` |
+| `last_updated` older than `contributed` | Set `last_updated` = `contributed` |
+| Name doesn't match filename | Update `name` field to match filename |
+
+Issues that CANNOT be auto-fixed:
+- Broken cross-references (need human decision on correct target)
+- Missing solution content (need domain knowledge)
+- Unclosed code blocks (need to identify which block is unclosed)
+
+---
+
+## Success Criteria
+
+- [ ] All files in scope are scanned
+- [ ] Frontmatter validation catches missing/invalid fields
+- [ ] Section validation catches missing required sections
+- [ ] Cross-reference validation catches broken links
+- [ ] Markdown syntax check catches unclosed blocks
+- [ ] Report clearly separates FAIL from WARN
+- [ ] Auto-fix offered for simple issues
+- [ ] Parallel execution used for full-library scans
+
+---
+
+## Related Commands
+
+- `/fire-add-new-skill` — Create new skills (run validate after)
+- `/fire-research` — Bulk skill creation (run validate after)
+- `/fire-search` — Search the validated library
+- `/fire-test` — Full plugin test suite (includes skill structure tests)
+
+---
+
+*Dominion Flow v12.8 — Trust but verify. Every skill, every field, every reference.*