By tracemem
Decision memory for Claude Code. Automatically tracks architectural decisions, code changes, and engineering choices into TraceMem's governed audit trail.
Close the active TraceMem decision envelope
Verify TraceMem plugin setup and connectivity
Add a note to the active TraceMem decision
Open a new TraceMem decision envelope for the current task
Get the immutable receipt for a closed decision
Reviews the accumulated decision trace at the end of a session, identifies gaps, and ensures the decision record is complete before closing. Reviews decision traces for completeness, identifies significant decisions that may have been missed, generates decision summaries for team visibility, and suggests additional context before closing a decision envelope.
Specialized agent that identifies and records key architectural and engineering decisions during coding sessions. Identifies architectural decisions during code changes, records technology choices with rationale and alternatives considered, captures schema changes, API contract modifications, and infrastructure decisions, adds structured context to TraceMem decision envelopes, and uses one-shot decision recording when no envelope is active.
Executes bash commands
Hook triggers when Bash tool is used
Modifies files
Hook triggers on file write and edit operations
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Uses power tools
Uses Bash, Write, or Edit tools
Uses power tools
Uses Bash, Write, or Edit tools
Decision memory for Claude Code. Automatically tracks architectural decisions, code changes, and engineering choices into TraceMem's governed audit trail.
From GitHub (recommended):
/plugin marketplace add tracemem/tracemem-claude-code-plugin
/plugin install tracemem@tracemem-plugins
Or install directly from GitHub without adding the marketplace:
claude plugin install tracemem@tracemem-plugins --marketplace-source github:tracemem/tracemem-claude-code-plugin
From a local clone:
git clone https://github.com/tracemem/tracemem-claude-code-plugin.git
Then in Claude Code:
/plugin install /path/to/tracemem-claude-code-plugin
# Add to ~/.bashrc, ~/.zshrc, or your shell profile
export TRACEMEM_API_KEY=your_api_key_here
Or in your project's .env file:
TRACEMEM_API_KEY=your_api_key_here
Get your API key from app.tracemem.com under your project settings.
/tracemem-doctor
This checks your API key, MCP server connectivity, and available tools.
| Command | Description |
|---|---|
/tracemem-open [action] | Open a decision envelope for the current task |
/tracemem-close [commit|abort] | Close the active decision envelope |
/tracemem-note <message> | Add a note to the active decision |
/tracemem-status | Show the active decision state |
/tracemem-trace [id] | View the decision trace |
/tracemem-receipt [id] | Get the immutable receipt for a closed decision |
/tracemem-record <description> | Record a one-shot architecture/engineering decision |
/tracemem-search [query] | Search past decisions for precedent |
/tracemem-doctor | Verify plugin setup and connectivity |
/tracemem-open| Action | Intent | Use when |
|---|---|---|
edit_files | code.change.apply | General code editing (default) |
refactor | code.refactor.execute | Restructuring code |
run_command | ops.command.execute | Operational commands |
deploy | deploy.release.execute | Deployments and releases |
secrets | secrets.change.propose | Secret modifications |
db_change | data.change.apply | Database changes |
review | code.review.assist | Code review |
The plugin uses three layers to ensure decisions are captured:
The decision-tracking skill teaches Claude when and how to track decisions autonomously. When Claude detects significant technical choices (architecture, dependencies, schema changes, API modifications), it records them automatically.
Lifecycle hooks provide guaranteed automation:
Two specialized agents handle decision workflows:
/tracemem-record)For architecture and engineering decisions that don't require data product operations. This is the most common path during coding:
/tracemem-record Use Redis for session caching
Claude will search for existing decisions, gather details from context, and record the decision in a single operation.
/tracemem-open ... /tracemem-close)For decisions involving governed data reads/writes, policy evaluations, or approval workflows:
/tracemem-open edit_files
... work on the task ...
/tracemem-close
| Variable | Required | Description |
|---|---|---|
TRACEMEM_API_KEY | Yes | API key for TraceMem authentication |
TRACEMEM_MCP_URL | No | Override MCP server URL (default: https://mcp.tracemem.com) |
The plugin connects to TraceMem's cloud-hosted MCP server via HTTP. No local process to manage. The connection is configured in .mcp.json and started automatically when the plugin is enabled.
Always tracked:
Never tracked:
npx claudepluginhub tracemem/tracemem-claude-code-plugin --plugin tracememLogs decisions and problem-solving approaches to survive context compaction
Long-term memory for Claude Code. Recalls architecture, conventions, and decisions across sessions — so Claude explains your codebase without file traversal, follows your patterns, and never re-asks what it already learned.
Curated persistent memory for Claude Code. Write gate prevents bloat — only behavior-changing facts get saved. Tiered architecture: daily logs, structured registers, and auto-loaded working memory.
Capture AI coding sessions to understand intent
Persistent memory across Claude Code sessions using Cognis
Persistent memory across Claude Code sessions using Supermemory