You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Detect and warn when new decisions or implementations contradict earlier documented decisions.
Problem
Over time, developers make architectural decisions that get documented as observations. Later, they might:
Forget about earlier decisions
Accidentally implement something that contradicts a previous choice
Not realize a new pattern conflicts with established conventions
Currently, claude-mem doesn't detect these conflicts.
Proposed Solution
1. Decision Tracking Enhancement
Tag observations with decision categories:
Architecture patterns (e.g., "use repository pattern", "avoid global state")
Technology choices (e.g., "use PostgreSQL", "prefer REST over GraphQL")
Code conventions (e.g., "no abbreviations in variable names")
Security policies (e.g., "always validate input at boundaries")
2. Conflict Detection Engine
When a new observation is created, compare against existing decisions:
New observation: "Added global state manager for user preferences"
⚠️ Potential conflict detected:
Decision from 2025-11-15: "Avoid global state, use dependency injection"
Session: "Refactoring auth module"
Options:
- This supersedes the old decision
- This is an exception (document why)
- Cancel and reconsider approach
3. Semantic Similarity
Use vector embeddings to detect semantic conflicts, not just keyword matches:
"Use composition over inheritance" vs "Extended BaseController class"
"Keep functions pure" vs "Added side effect to calculate total"
4. Conflict Resolution
Store resolution metadata:
{
"observation_id": 456,
"conflicts_with": 123,
"resolution": "supersedes",
"reason": "Requirements changed, global state now acceptable for this use case"
}
Use Cases
Architectural Consistency: Catch drift from established patterns
Team Alignment: Ensure decisions are followed across sessions
Decision History: Track why decisions evolved over time
Code Review Aid: Surface conflicts before they become tech debt
Implementation Notes
Could use Chroma embeddings for semantic comparison
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
Summary
Detect and warn when new decisions or implementations contradict earlier documented decisions.
Problem
Over time, developers make architectural decisions that get documented as observations. Later, they might:
Currently, claude-mem doesn't detect these conflicts.
Proposed Solution
1. Decision Tracking Enhancement
Tag observations with decision categories:
2. Conflict Detection Engine
When a new observation is created, compare against existing decisions:
3. Semantic Similarity
Use vector embeddings to detect semantic conflicts, not just keyword matches:
4. Conflict Resolution
Store resolution metadata:
{ "observation_id": 456, "conflicts_with": 123, "resolution": "supersedes", "reason": "Requirements changed, global state now acceptable for this use case" }Use Cases
Implementation Notes
🤖 Generated with Claude Code
Beta Was this translation helpful? Give feedback.
All reactions