Analyze and fix project alignment with PROJECT.md using GenAI
/plugin marketplace add akaszubski/autonomous-dev/plugin install autonomous-dev@autonomous-devInvoke the alignment-analyzer agent to find conflicts between PROJECT.md and reality, then guide interactive fixing.
GenAI-powered project alignment analysis and interactive fixing
/align-project
Time: 5-20 minutes (depending on issues found) Interactive: Agent asks about each conflict found GenAI-Powered: Uses alignment-analyzer agent for intelligent analysis
The alignment-analyzer agent works with you to align your project:
The agent:
For each conflict found, the agent asks:
PROJECT.md says: "REST API only, no GraphQL"
Reality shows: graphql/ directory with resolvers
Status: GraphQL implementation exists
What should we do?
A) YES - PROJECT.md is correct (remove GraphQL)
B) NO - Update PROJECT.md to include GraphQL
Your choice [A/B]:
The agent intelligently:
After all questions answered, the agent:
🔍 Analyzing project alignment with PROJECT.md...
Found 5 conflicts:
CONFLICT #1: GraphQL Scope Mismatch
────────────────────────────────────
PROJECT.md says:
SCOPE (Out of Scope):
- GraphQL API
Reality shows:
graphql/ directory exists with resolvers implemented
What should we do?
A) YES - PROJECT.md is correct (remove GraphQL code)
B) NO - Update PROJECT.md to include GraphQL
Your choice [A/B]: A
Action: Remove graphql/ directory and update PROJECT.md
Impact: Removes ~200 lines of code (not in scope)
CONFLICT #2: Missing Test Structure
────────────────────────────────────
PROJECT.md says:
Tests organized in: tests/unit/, tests/integration/
Reality shows:
Only tests/ directory exists (no subdirectories)
What should we do?
A) YES - PROJECT.md is correct (organize tests)
B) NO - Update PROJECT.md to reflect current structure
Your choice [A/B]: A
Action: Organize existing tests into unit/integration/uat
Impact: Makes testing structure match documentation
# 1. Run alignment
/align-project
# 2. Agent analyzes and asks about each conflict
# 3. You choose A (PROJECT.md correct) or B (update PROJECT.md)
# 4. Agent implements your choices
# 5. Agent commits changes with summary
# Check alignment
/align-project
# If fully aligned:
# ✅ No conflicts found - project is well-aligned!
# If conflicts found:
# Answer each question to resolve
✅ Analysis before action: Agent shows conflicts before making changes ✅ Binary questions: No ambiguity - clear choices ✅ Clear impact: Agent explains what each choice does ✅ Reversible: Changes are committed (easy rollback) ✅ Stop anytime: Can exit at any point
Run /align-project when:
No need if:
# Create PROJECT.md first
/setup
# Then run alignment
/align-project
Agent analyzes based on PROJECT.md being the source of truth. If you disagree:
B) NO to update PROJECT.md instead/setup - Create or update PROJECT.md/test - Run tests after alignment/auto-implement - Autonomous feature developmentUse this to find conflicts between your documented goals and actual implementation. Agent intelligently guides you to alignment.