specflow-cc 1.6.3 → 1.7.0

package/CHANGELOG.md

@@ -5,6 +5,33 @@ All notable changes to SpecFlow will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [1.7.0] - 2026-01-28
+
+### Added
+
+- **AI pre-analysis of external feedback** — `/sf:revise` now analyzes imported external audit items before showing review options
+  - Provides recommendations: ✓ Apply / ? Discuss / ✗ Skip with reasoning
+  - New "Apply recommended" option to apply only recommended items
+  - `--no-analysis` flag to skip pre-analysis and go directly to manual review
+
+### Changed
+
+- `/sf:audit --import` workflow now flows through AI analysis step by default
+- Updated help documentation with new flag
+
+---
+
+## [1.6.3] - 2026-01-28
+
+### Added
+
+- `/sf:audit --import "feedback"` — Import external feedback (code reviews, security audits) for critical evaluation
+- `/sf:revise` detects external audits and offers per-item evaluation (Apply/Skip/Discuss/Defer)
+- New `external_review` status in state machine
+- Strategic sanity check in auditor agents
+
+---
+
 ## [1.6.0] - 2026-01-25
 
 ### Added

package/README.md

@@ -339,9 +339,14 @@ Six months later, you can read the spec and understand not just *what* was built
 
 ```bash
 /sf:audit --import "[Critical] SQL injection; [Major] Add rate limiting"
-/sf:revise              # Critically evaluate each item (Apply/Skip/Defer)
+/sf:revise              # AI analyzes feedback, then shows review options
+/sf:revise --no-analysis  # Skip AI analysis, go directly to manual review
 ```
 
+When reviewing external feedback, Claude first analyzes each item against the
+specification and provides recommendations (✓ Apply / ? Discuss / ✗ Skip)
+before showing interactive options.
+
 ### Research & Clarification
 
 | Command | Description |

package/commands/sf/help.md

@@ -148,11 +148,22 @@ Read the command file and extract:
 ## Examples
 
 ```
-/sf:revise                 # Interactive — shows audit comments
+/sf:revise                 # Interactive — shows audit comments with analysis
 /sf:revise all             # Apply all audit comments
 /sf:revise 1,2             # Apply only comments 1 and 2
+/sf:revise --no-analysis   # Skip pre-analysis, go directly to review mode
 /sf:revise "add error handling, ignore item 3"
 ```
+
+## Flags
+
+| Flag | Description |
+|------|-------------|
+| `--no-analysis` | Skip AI pre-analysis of external feedback |
+
+**Note:** For external audits, Claude analyzes each feedback item and provides
+recommendations (Apply/Discuss/Skip) before showing review options. Use
+`--no-analysis` to skip this and go directly to manual review.
 ```
 
 **For `/sf:split`:**

package/commands/sf/revise.md

@@ -86,6 +86,65 @@ else
 fi
 ```
 
+## Step 4.5: Pre-analyze External Feedback
+
+**Only for external audits (AUDIT_TYPE="external"):**
+
+Skip this step if:
+- `--no-analysis` flag is present in arguments
+- AUDIT_TYPE is "internal"
+
+### Analysis Process
+
+Read the specification and project context, then analyze each external feedback item:
+
+For each item in the external audit:
+1. Check if the issue actually exists in the current specification
+2. Evaluate relevance to the specification's scope and goals
+3. Consider project architecture and constraints
+4. Determine recommendation: Apply / Discuss / Skip
+
+### Display Analysis Results
+
+```
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ ANALYSIS: EXTERNAL FEEDBACK — SPEC-XXX
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+Based on specification context and project architecture:
+
+| # | Item | Recommendation | Reason |
+|---|------|----------------|--------|
+| 1 | {issue description} | ✓ Apply | {why it's relevant} |
+| 2 | {issue description} | ✓ Apply | {why it's relevant} |
+| 3 | {issue description} | ? Discuss | {why clarification needed} |
+| 4 | {issue description} | ✗ Skip | {why not applicable} |
+
+**Summary:** {N} recommended, {M} need discussion, {K} likely irrelevant
+
+---
+```
+
+### Recommendation Criteria
+
+**✓ Apply** — recommend when:
+- Issue clearly exists in the specification
+- Fix aligns with specification goals
+- Within scope of current work
+
+**? Discuss** — recommend when:
+- Issue may be valid but requires clarification
+- Trade-offs need user decision
+- Significant architectural change implied
+
+**✗ Skip** — recommend when:
+- Issue doesn't exist in current specification
+- Clearly out of scope for this specification
+- Based on incorrect assumptions about the project
+- Would conflict with explicit requirements
+
+Continue to Step 5 with analysis context available.
+
 ## Step 5: Parse Arguments
 
 | Argument | Action |
@@ -93,8 +152,18 @@ fi
 | (none) | Interactive mode — show comments, ask what to fix |
 | "all" | Apply all critical issues AND recommendations |
 | "1,2,3" | Apply only numbered items |
+| "--no-analysis" | Skip pre-analysis, go directly to review mode |
 | "..." | Treat as custom revision instructions |
 
+**Check for `--no-analysis` flag:**
+```bash
+if echo "$ARGS" | grep -q "\-\-no-analysis"; then
+    SKIP_ANALYSIS=true
+    # Remove flag from args for further processing
+    ARGS=$(echo "$ARGS" | sed 's/--no-analysis//g' | xargs)
+fi
+```
+
 ### If Interactive Mode (no arguments):
 
 Display audit comments with context about source:
@@ -128,6 +197,38 @@ Use AskUserQuestion with options:
 
 **For external audit (requires critical evaluation):**
 
+**If pre-analysis was performed (Step 4.5):**
+
+```
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ REVIEW EXTERNAL FEEDBACK: SPEC-XXX
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+
+External Audit (v{N}) — with analysis:
+
+| # | Item | Severity | Recommendation | Reason |
+|---|------|----------|----------------|--------|
+| 1 | {issue} | Critical | ✓ Apply | {reason} |
+| 2 | {issue} | Critical | ✓ Apply | {reason} |
+| 3 | {issue} | Major | ? Discuss | {reason} |
+| 4 | {issue} | Major | ✗ Skip | {reason} |
+| 5 | {issue} | Minor | ✓ Apply | {reason} |
+
+**Analysis:** {N} recommended, {M} need discussion, {K} skip suggested
+
+---
+
+How to proceed?
+```
+
+Use AskUserQuestion with options:
+- "Apply recommended" → apply items marked ✓ Apply (recommended)
+- "Review each item" → interactive per-item evaluation
+- "Apply all" → apply everything (ignore analysis)
+- "Select specific" → ask for numbers
+
+**If pre-analysis was skipped (`--no-analysis` or internal flow):**
+
 ```
 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  REVIEW EXTERNAL FEEDBACK: SPEC-XXX

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "specflow-cc",
-  "version": "1.6.3",
+  "version": "1.7.0",
   "description": "Spec-driven development system for Claude Code — quality-first workflow with explicit audit cycles",
   "bin": {
     "specflow-cc": "bin/install.js"