runtime-memory 3.0.0__tar.gz → 4.0.0__tar.gz

{runtime_memory-3.0.0 → runtime_memory-4.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: runtime-memory
-Version: 3.0.0
+Version: 4.0.0
 Summary: Persistent memory for AI coding agents with outcome-based learning
 Project-URL: Homepage, https://github.com/runtimenoteslabs/memory-layer
 Project-URL: Documentation, https://github.com/runtimenoteslabs/memory-layer#readme
@@ -66,20 +66,32 @@ Persistent memory for AI coding agents with outcome-based learning.
 
 > **New to Runtime Memory?** See the [User Guide](https://github.com/runtimenoteslabs/memory-layer/blob/main/USER_GUIDE.md) for an introduction to using Runtime Memory with Claude Code.
 
-## What It Does
+## What it does
 
-Runtime Memory stores knowledge from your coding sessions and learns which memories actually help. When advice works, it gets boosted (+0.2). When it fails, it gets penalized (-0.3). Over time, good memories rise to the top.
+Runtime Memory stores knowledge from your coding sessions and records how often each memory worked and how often it failed. Among the memories relevant to a query, those with a better record rank higher. A memory that keeps failing is left out of retrieval until its failures fade.
 
 ## Installation
 
 ```bash
-pip install runtime-memory
+pip install "runtime-memory[all]"
 ```
 
+The base package stores and searches memories by keyword. Each extra adds a part:
+
+| Extra | Adds | Without it |
+|-------|------|------------|
+| `embedding` | Semantic search, with `sentence-transformers` | Search is keyword-only |
+| `extraction` | Extracting memories from sessions with Claude | No extraction |
+| `server` | The MCP server, the REST API and the web UI | `mem serve` is unavailable |
+| `all` | All three | |
+
+`mem stats` shows which search mode is in use. With `embedding`, the first search
+downloads an embedding model of about 100 MB and caches it.
+
 Or from source:
 
 ```bash
-pip install git+https://github.com/runtimenoteslabs/memory-layer.git
+pip install "runtime-memory[all] @ git+https://github.com/runtimenoteslabs/memory-layer.git"
 ```
 
 For development:
@@ -95,9 +107,7 @@ repository is still named memory-layer, which is where the project started; the
 package was renamed in 3.0. An unrelated package holds `memory-layer` on PyPI,
 so `pip install memory-layer` fetches that one instead of this project.
 
-**Note:** First run downloads an embedding model (~100MB) for semantic search. This happens once and is cached. Subsequent operations are fast (<100ms).
-
-## Quick Start
+## Quick start
 
 ### Python SDK
 
@@ -121,7 +131,7 @@ async with MemoryClient() as client:
     context = await client.get_context()
 ```
 
-### Synchronous Client
+### Synchronous client
 
 ```python
 from runtime_memory.sdk import SyncMemoryClient
@@ -140,6 +150,12 @@ mem add "Use type hints for better IDE support" -c convention
 # Search memories
 mem search "type hints"
 
+# See why a search returns what it does, and why the rest were left out
+mem why "type hints"
+
+# Store statistics: outcome records, search mode, and the Hermes trace if present
+mem stats
+
 # Record outcome
 mem outcome <memory-id> worked
 
@@ -170,7 +186,7 @@ curl -X POST http://localhost:8080/memories/search \
   -d '{"query": "testing"}'
 ```
 
-### MCP Server
+### MCP server
 
 For multi-agent setups, Runtime Memory provides an MCP server:
 
@@ -189,7 +205,7 @@ Configure in your MCP client:
 }
 ```
 
-#### Multi-Agent Configurations
+#### Multi-agent configurations
 
 All agents share the same memory store. Memories created in Claude Code appear in Cursor, feedback from OpenCode improves results everywhere.
 
@@ -229,14 +245,14 @@ All agents share the same memory store. Memories created in Claude Code appear i
 }
 ```
 
-### Claude Code Integration
+### Claude Code integration
 
 Runtime Memory integrates with Claude Code via hooks and skills. For a beginner-friendly walkthrough, see the [User Guide](https://github.com/runtimenoteslabs/memory-layer/blob/main/USER_GUIDE.md).
 
 **Installation:**
 
 ```bash
-pip install runtime-memory
+pip install "runtime-memory[all]"
 
 # Go to your project directory
 cd your-project
@@ -275,7 +291,7 @@ The `mem install-plugin` command creates:
 /memory-context                  # Get project context
 ```
 
-### Task Integration (Beads + Claude Code)
+### Task integration (Beads and Claude Code)
 
 Runtime Memory integrates with task trackers to automatically learn from task outcomes.
 
@@ -286,7 +302,7 @@ Runtime Memory integrates with task trackers to automatically learn from task ou
 **How it works:**
 1. You work on a task, Claude searches for relevant memories
 2. Those memories get linked to your task
-3. When you mark the task done, linked memories are automatically boosted
+3. When you mark the task done, the linked memories are recorded as having worked
 
 ```bash
 # Unified task commands (all sources)
@@ -303,25 +319,27 @@ mem beads-context
 mem beads-stats
 ```
 
-No setup required - Runtime Memory auto-detects both `.beads/` and `~/.claude/todos/` directories.
+Runtime Memory finds `.beads/` and `~/.claude/todos/` itself; there is nothing to configure.
 
 **Environment variables:**
 - `CLAUDE_CODE_TASK_LIST_ID` - Filter to specific task list
 - `CLAUDE_CODE_TODOS_DIR` - Custom todos directory location
 
-### Hermes Agent Integration
+### Hermes Agent integration
 
 Runtime Memory can serve as Hermes Agent's memory provider, replacing its capped
 note file with retrieval over the same store Claude Code and MCP clients use.
 
 ```bash
 # Install into the environment Hermes runs in
-~/.hermes/hermes-agent/venv/bin/python -m pip install \
-    git+https://github.com/runtimenoteslabs/memory-layer.git
+~/.hermes/hermes-agent/venv/bin/python -m pip install "runtime-memory[embedding,extraction]"
 
 hermes config set memory.provider runtimememory
 ```
 
+Without the `embedding` extra in Hermes' own environment, the provider searches by
+keyword only and logs a warning that names the interpreter.
+
 Hermes finds the provider through the `hermes_agent.memory_providers` entry
 point, so you do not edit its code or config files by hand. See
 [docs/hermes.md](https://github.com/runtimenoteslabs/memory-layer/blob/main/docs/hermes.md) for configuration, the tool surface, and the
@@ -347,9 +365,9 @@ mem serve --rest --port 8080
 - Record outcomes
 - Light/dark theme
 
-## Memory Categories
+## Memory categories
 
-| Category | Use For | Example |
+| Category | Use for | Example |
 |----------|---------|---------|
 | `architecture` | System design | "Microservices with event sourcing" |
 | `convention` | Coding standards | "Use snake_case for Python" |
@@ -361,42 +379,71 @@ mem serve --rest --port 8080
 | `command` | Useful commands | "npm run test:coverage" |
 | `preference` | User preferences | "Prefer functional style" |
 
-## Outcome Scoring
+## Outcome scoring
 
-| Outcome | Score Change | When to Use |
-|---------|--------------|-------------|
-| `worked` | +0.2 | Advice solved the problem |
-| `failed` | -0.3 | Advice was wrong or unhelpful |
-| `partial` | +0.05 | Advice was on the right track |
+| Outcome | Adds | When to use |
+|---------|------|-------------|
+| `worked` | one success | Advice solved the problem |
+| `failed` | one failure | Advice was wrong or unhelpful |
+| `partial` | a quarter of a success | Advice was on the right track |
 
-The asymmetric scoring is intentional: bad advice wastes debugging time and erodes trust, so it's penalized more heavily.
+A memory's outcome score is `(worked - 1.5 x failed) / (worked + 1.5 x failed + 2)`,
+between -1 and 1. One success gives 0.33 and ten give 0.83, so a single
+observation counts for less than a long record. A failure weighs 1.5 successes,
+because following bad advice wastes debugging time. Each count halves every 90
+days.
 
-## How Retrieval Works
+Retrieval leaves out a memory whose score is -0.5 or lower, which takes two
+failures and no successes. One failure is not enough, because it may have been
+blamed on the wrong memory. The memory is retrieved again once its failures
+have faded.
 
-Runtime Memory uses a 5-signal hybrid retrieval system that combines multiple relevance signals:
+To change these values, see `RetrievalConfig.outcome_model` and
+`RetrievalConfig.failure_gate`.
+
+## How retrieval works
+
+Retrieval runs in two stages. Relevance to your query decides which memories
+compete, then the other signals order them.
+
+**Stage 1, the relevance pool.** A search keeps the `ceil(limit x 2)` memories
+most relevant to the query and drops any with no relevance at all. Outcome records
+reorder only the memories that match the query.
+
+**Stage 2, the score.**
 
 | Signal | Weight | Description |
 |--------|--------|-------------|
-| Semantic | 35% | Vector similarity to your query |
+| Semantic | 55% | Vector and keyword similarity to your query |
 | Outcome | 25% | Learned effectiveness from feedback |
-| Recency | 15% | Recent memories weighted higher (30-day half-life) |
-| Frequency | 15% | Frequently used memories rise |
 | Confidence | 10% | Extraction confidence score |
+| Recency | 10% | Newer memories weighted higher (30-day half-life on age) |
+| Frequency | 0% | Off by default; see below |
 
-Two of the five signals, outcome and frequency, come from how memories have
-performed rather than from the query, so ranking changes as feedback accumulates.
+Outcome and confidence come from how memories have performed rather than from
+the query, so ranking changes as feedback accumulates.
 
-### Category Boosting
+**What changed in 4.0.0, and why.** Tier 2 evaluation runs found the older
+scoring deciding retrieval on signals that had nothing to do with the query:
 
-When you ask about errors, troubleshooting memories get a 1.5x boost. Query intent is detected and the right category is prioritized:
+- **Frequency left the default score.** It rewards having been retrieved, which
+  is not evidence of having helped, and it compounds: a wrong memory held a top
+  place through a whole task sequence on it. Set `frequency_weight` to bring it
+  back.
+- **Category boosts are neutral.** Multiplying the whole score by a category
+  seated a memory that ranked about 25th on relevance at rank 1, and in another
+  run kept the one memory that would have prevented a repeated mistake out of
+  every prompt. Pass `category_boosts` to set your own.
+- **Recency decays from a memory's age,** not from when it was last touched.
+  Retrieval no longer moves that clock.
 
-| Query Pattern | Boosted Category | Multiplier |
-|---------------|------------------|------------|
-| "What went wrong..." | troubleshooting | 1.5x |
-| "Watch out for..." | gotcha | 1.4x |
-| "Why did we choose..." | decision | 1.4x |
-| "How should I structure..." | pattern, convention | 1.3x |
-| "System design..." | architecture | 1.2x |
+`RetrievalConfig.legacy_3x()` restores the 3.x weights, boosts and single-stage
+scoring if you tuned for them.
+
+### Category routing
+
+`CategoryRouter` maps query wording to a category, but no search path calls it.
+It is available to callers that want to pass `category=` themselves.
 
 ## Results
 
@@ -411,25 +458,29 @@ After 12 weeks of use:
 
 ## Configuration
 
-### Environment Variables
+### Environment variables
 
 | Variable | Description | Default |
 |----------|-------------|---------|
 | `ANTHROPIC_API_KEY` | For LLM-based extraction | Required for extraction features |
-| `MEMORY_LAYER_DB` | Database location | `~/.runtime-memory/memories.db` |
-| `MEMORY_LAYER_ENV` | Environment (development/testing/production) | development |
-| `MEMORY_LAYER_LOG_LEVEL` | Logging level | WARNING |
+| `RUNTIME_MEMORY_DB` | Database location, read by the CLI, the MCP server and the Hermes provider | `~/.runtime-memory/memories.db` |
+| `RUNTIME_MEMORY_ENV` | Environment (development/testing/production) | development |
+| `RUNTIME_MEMORY_LOG_LEVEL` | Logging level | WARNING |
 | `CLAUDE_CODE_TASK_LIST_ID` | Filter Claude Code tasks | None |
 | `CLAUDE_CODE_TODOS_DIR` | Custom todos directory | `~/.claude/todos/` |
 
-### Data Location
+Variables set with the pre-3.0 prefix `MEMORY_LAYER_` are still read, under their
+`RUNTIME_MEMORY_` names. The Hermes provider's own settings are in
+[docs/hermes.md](https://github.com/runtimenoteslabs/memory-layer/blob/main/docs/hermes.md).
+
+### Data location
 
 ```
 ~/.runtime-memory/
 └── memories.db    # SQLite database
 ```
 
-## Project Structure
+## Project structure
 
 ```
 memory-layer/

{runtime_memory-3.0.0 → runtime_memory-4.0.0}/README.md

@@ -4,20 +4,32 @@ Persistent memory for AI coding agents with outcome-based learning.
 
 > **New to Runtime Memory?** See the [User Guide](https://github.com/runtimenoteslabs/memory-layer/blob/main/USER_GUIDE.md) for an introduction to using Runtime Memory with Claude Code.
 
-## What It Does
+## What it does
 
-Runtime Memory stores knowledge from your coding sessions and learns which memories actually help. When advice works, it gets boosted (+0.2). When it fails, it gets penalized (-0.3). Over time, good memories rise to the top.
+Runtime Memory stores knowledge from your coding sessions and records how often each memory worked and how often it failed. Among the memories relevant to a query, those with a better record rank higher. A memory that keeps failing is left out of retrieval until its failures fade.
 
 ## Installation
 
 ```bash
-pip install runtime-memory
+pip install "runtime-memory[all]"
 ```
 
+The base package stores and searches memories by keyword. Each extra adds a part:
+
+| Extra | Adds | Without it |
+|-------|------|------------|
+| `embedding` | Semantic search, with `sentence-transformers` | Search is keyword-only |
+| `extraction` | Extracting memories from sessions with Claude | No extraction |
+| `server` | The MCP server, the REST API and the web UI | `mem serve` is unavailable |
+| `all` | All three | |
+
+`mem stats` shows which search mode is in use. With `embedding`, the first search
+downloads an embedding model of about 100 MB and caches it.
+
 Or from source:
 
 ```bash
-pip install git+https://github.com/runtimenoteslabs/memory-layer.git
+pip install "runtime-memory[all] @ git+https://github.com/runtimenoteslabs/memory-layer.git"
 ```
 
 For development:
@@ -33,9 +45,7 @@ repository is still named memory-layer, which is where the project started; the
 package was renamed in 3.0. An unrelated package holds `memory-layer` on PyPI,
 so `pip install memory-layer` fetches that one instead of this project.
 
-**Note:** First run downloads an embedding model (~100MB) for semantic search. This happens once and is cached. Subsequent operations are fast (<100ms).
-
-## Quick Start
+## Quick start
 
 ### Python SDK
 
@@ -59,7 +69,7 @@ async with MemoryClient() as client:
     context = await client.get_context()
 ```
 
-### Synchronous Client
+### Synchronous client
 
 ```python
 from runtime_memory.sdk import SyncMemoryClient
@@ -78,6 +88,12 @@ mem add "Use type hints for better IDE support" -c convention
 # Search memories
 mem search "type hints"
 
+# See why a search returns what it does, and why the rest were left out
+mem why "type hints"
+
+# Store statistics: outcome records, search mode, and the Hermes trace if present
+mem stats
+
 # Record outcome
 mem outcome <memory-id> worked
 
@@ -108,7 +124,7 @@ curl -X POST http://localhost:8080/memories/search \
   -d '{"query": "testing"}'
 ```
 
-### MCP Server
+### MCP server
 
 For multi-agent setups, Runtime Memory provides an MCP server:
 
@@ -127,7 +143,7 @@ Configure in your MCP client:
 }
 ```
 
-#### Multi-Agent Configurations
+#### Multi-agent configurations
 
 All agents share the same memory store. Memories created in Claude Code appear in Cursor, feedback from OpenCode improves results everywhere.
 
@@ -167,14 +183,14 @@ All agents share the same memory store. Memories created in Claude Code appear i
 }
 ```
 
-### Claude Code Integration
+### Claude Code integration
 
 Runtime Memory integrates with Claude Code via hooks and skills. For a beginner-friendly walkthrough, see the [User Guide](https://github.com/runtimenoteslabs/memory-layer/blob/main/USER_GUIDE.md).
 
 **Installation:**
 
 ```bash
-pip install runtime-memory
+pip install "runtime-memory[all]"
 
 # Go to your project directory
 cd your-project
@@ -213,7 +229,7 @@ The `mem install-plugin` command creates:
 /memory-context                  # Get project context
 ```
 
-### Task Integration (Beads + Claude Code)
+### Task integration (Beads and Claude Code)
 
 Runtime Memory integrates with task trackers to automatically learn from task outcomes.
 
@@ -224,7 +240,7 @@ Runtime Memory integrates with task trackers to automatically learn from task ou
 **How it works:**
 1. You work on a task, Claude searches for relevant memories
 2. Those memories get linked to your task
-3. When you mark the task done, linked memories are automatically boosted
+3. When you mark the task done, the linked memories are recorded as having worked
 
 ```bash
 # Unified task commands (all sources)
@@ -241,25 +257,27 @@ mem beads-context
 mem beads-stats
 ```
 
-No setup required - Runtime Memory auto-detects both `.beads/` and `~/.claude/todos/` directories.
+Runtime Memory finds `.beads/` and `~/.claude/todos/` itself; there is nothing to configure.
 
 **Environment variables:**
 - `CLAUDE_CODE_TASK_LIST_ID` - Filter to specific task list
 - `CLAUDE_CODE_TODOS_DIR` - Custom todos directory location
 
-### Hermes Agent Integration
+### Hermes Agent integration
 
 Runtime Memory can serve as Hermes Agent's memory provider, replacing its capped
 note file with retrieval over the same store Claude Code and MCP clients use.
 
 ```bash
 # Install into the environment Hermes runs in
-~/.hermes/hermes-agent/venv/bin/python -m pip install \
-    git+https://github.com/runtimenoteslabs/memory-layer.git
+~/.hermes/hermes-agent/venv/bin/python -m pip install "runtime-memory[embedding,extraction]"
 
 hermes config set memory.provider runtimememory
 ```
 
+Without the `embedding` extra in Hermes' own environment, the provider searches by
+keyword only and logs a warning that names the interpreter.
+
 Hermes finds the provider through the `hermes_agent.memory_providers` entry
 point, so you do not edit its code or config files by hand. See
 [docs/hermes.md](https://github.com/runtimenoteslabs/memory-layer/blob/main/docs/hermes.md) for configuration, the tool surface, and the
@@ -285,9 +303,9 @@ mem serve --rest --port 8080
 - Record outcomes
 - Light/dark theme
 
-## Memory Categories
+## Memory categories
 
-| Category | Use For | Example |
+| Category | Use for | Example |
 |----------|---------|---------|
 | `architecture` | System design | "Microservices with event sourcing" |
 | `convention` | Coding standards | "Use snake_case for Python" |
@@ -299,42 +317,71 @@ mem serve --rest --port 8080
 | `command` | Useful commands | "npm run test:coverage" |
 | `preference` | User preferences | "Prefer functional style" |
 
-## Outcome Scoring
+## Outcome scoring
 
-| Outcome | Score Change | When to Use |
-|---------|--------------|-------------|
-| `worked` | +0.2 | Advice solved the problem |
-| `failed` | -0.3 | Advice was wrong or unhelpful |
-| `partial` | +0.05 | Advice was on the right track |
+| Outcome | Adds | When to use |
+|---------|------|-------------|
+| `worked` | one success | Advice solved the problem |
+| `failed` | one failure | Advice was wrong or unhelpful |
+| `partial` | a quarter of a success | Advice was on the right track |
 
-The asymmetric scoring is intentional: bad advice wastes debugging time and erodes trust, so it's penalized more heavily.
+A memory's outcome score is `(worked - 1.5 x failed) / (worked + 1.5 x failed + 2)`,
+between -1 and 1. One success gives 0.33 and ten give 0.83, so a single
+observation counts for less than a long record. A failure weighs 1.5 successes,
+because following bad advice wastes debugging time. Each count halves every 90
+days.
 
-## How Retrieval Works
+Retrieval leaves out a memory whose score is -0.5 or lower, which takes two
+failures and no successes. One failure is not enough, because it may have been
+blamed on the wrong memory. The memory is retrieved again once its failures
+have faded.
 
-Runtime Memory uses a 5-signal hybrid retrieval system that combines multiple relevance signals:
+To change these values, see `RetrievalConfig.outcome_model` and
+`RetrievalConfig.failure_gate`.
+
+## How retrieval works
+
+Retrieval runs in two stages. Relevance to your query decides which memories
+compete, then the other signals order them.
+
+**Stage 1, the relevance pool.** A search keeps the `ceil(limit x 2)` memories
+most relevant to the query and drops any with no relevance at all. Outcome records
+reorder only the memories that match the query.
+
+**Stage 2, the score.**
 
 | Signal | Weight | Description |
 |--------|--------|-------------|
-| Semantic | 35% | Vector similarity to your query |
+| Semantic | 55% | Vector and keyword similarity to your query |
 | Outcome | 25% | Learned effectiveness from feedback |
-| Recency | 15% | Recent memories weighted higher (30-day half-life) |
-| Frequency | 15% | Frequently used memories rise |
 | Confidence | 10% | Extraction confidence score |
+| Recency | 10% | Newer memories weighted higher (30-day half-life on age) |
+| Frequency | 0% | Off by default; see below |
 
-Two of the five signals, outcome and frequency, come from how memories have
-performed rather than from the query, so ranking changes as feedback accumulates.
+Outcome and confidence come from how memories have performed rather than from
+the query, so ranking changes as feedback accumulates.
 
-### Category Boosting
+**What changed in 4.0.0, and why.** Tier 2 evaluation runs found the older
+scoring deciding retrieval on signals that had nothing to do with the query:
 
-When you ask about errors, troubleshooting memories get a 1.5x boost. Query intent is detected and the right category is prioritized:
+- **Frequency left the default score.** It rewards having been retrieved, which
+  is not evidence of having helped, and it compounds: a wrong memory held a top
+  place through a whole task sequence on it. Set `frequency_weight` to bring it
+  back.
+- **Category boosts are neutral.** Multiplying the whole score by a category
+  seated a memory that ranked about 25th on relevance at rank 1, and in another
+  run kept the one memory that would have prevented a repeated mistake out of
+  every prompt. Pass `category_boosts` to set your own.
+- **Recency decays from a memory's age,** not from when it was last touched.
+  Retrieval no longer moves that clock.
 
-| Query Pattern | Boosted Category | Multiplier |
-|---------------|------------------|------------|
-| "What went wrong..." | troubleshooting | 1.5x |
-| "Watch out for..." | gotcha | 1.4x |
-| "Why did we choose..." | decision | 1.4x |
-| "How should I structure..." | pattern, convention | 1.3x |
-| "System design..." | architecture | 1.2x |
+`RetrievalConfig.legacy_3x()` restores the 3.x weights, boosts and single-stage
+scoring if you tuned for them.
+
+### Category routing
+
+`CategoryRouter` maps query wording to a category, but no search path calls it.
+It is available to callers that want to pass `category=` themselves.
 
 ## Results
 
@@ -349,25 +396,29 @@ After 12 weeks of use:
 
 ## Configuration
 
-### Environment Variables
+### Environment variables
 
 | Variable | Description | Default |
 |----------|-------------|---------|
 | `ANTHROPIC_API_KEY` | For LLM-based extraction | Required for extraction features |
-| `MEMORY_LAYER_DB` | Database location | `~/.runtime-memory/memories.db` |
-| `MEMORY_LAYER_ENV` | Environment (development/testing/production) | development |
-| `MEMORY_LAYER_LOG_LEVEL` | Logging level | WARNING |
+| `RUNTIME_MEMORY_DB` | Database location, read by the CLI, the MCP server and the Hermes provider | `~/.runtime-memory/memories.db` |
+| `RUNTIME_MEMORY_ENV` | Environment (development/testing/production) | development |
+| `RUNTIME_MEMORY_LOG_LEVEL` | Logging level | WARNING |
 | `CLAUDE_CODE_TASK_LIST_ID` | Filter Claude Code tasks | None |
 | `CLAUDE_CODE_TODOS_DIR` | Custom todos directory | `~/.claude/todos/` |
 
-### Data Location
+Variables set with the pre-3.0 prefix `MEMORY_LAYER_` are still read, under their
+`RUNTIME_MEMORY_` names. The Hermes provider's own settings are in
+[docs/hermes.md](https://github.com/runtimenoteslabs/memory-layer/blob/main/docs/hermes.md).
+
+### Data location
 
 ```
 ~/.runtime-memory/
 └── memories.db    # SQLite database
 ```
 
-## Project Structure
+## Project structure
 
 ```
 memory-layer/

{runtime_memory-3.0.0 → runtime_memory-4.0.0}/src/runtime_memory/__init__.py

@@ -10,7 +10,7 @@ This package provides:
 
 from __future__ import annotations
 
-__version__ = "3.0.0"
+__version__ = "4.0.0"
 __author__ = "exitcode42"
 
 from runtime_memory.core.legacy_env import apply_legacy_env

{runtime_memory-3.0.0 → runtime_memory-4.0.0}/src/runtime_memory/claude_code/commands.py

@@ -373,13 +373,16 @@ class CommandHandler:
                 error=f"No memory found with ID: {memory_id}",
             )
 
-        # Record the outcome
+        # Record the outcome. The engine stores no notes, so they are only echoed back.
         old_score = memory.outcome_score
-        updated_memory = await self.engine.record_outcome(
-            memory_id=memory_id,
-            outcome=outcome,
-            context=notes,
-        )
+        updated = await self.engine.record_outcome([memory_id], outcome)
+        if not updated:
+            return CommandResult(
+                success=False,
+                command=CommandType.OUTCOME,
+                message="Memory not found",
+                error=f"No memory found with ID: {memory_id}",
+            )
 
         return CommandResult(
             success=True,
@@ -389,7 +392,7 @@ class CommandHandler:
                 "memory_id": memory_id,
                 "outcome": outcome.value,
                 "old_score": old_score,
-                "new_score": updated_memory.outcome_score,
+                "new_score": updated[0].outcome_score,
                 "notes": notes,
             },
         )