Conversation
Native Bub integration via nowledge-mem-bub package: - 9 tools (mem.search, mem.save, mem.context, mem.connections, etc.) - 3 hooks (system_prompt, load_state, save_state) - Bundled nowledge-mem skill for agent self-guidance - Two modes: default (agent-driven) + session context (auto-inject) - Conversations captured to Nowledge Mem threads for cross-tool discovery - Access Anywhere support via config file or env vars Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> chore: remove .pyc files and add .gitignore for Bub plugin Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
📝 WalkthroughWalkthroughAdds a new nowledge-mem-bub-plugin: packaging and docs, an async nmem CLI client (NmemClient/NmemError), a Bub plugin (NowledgeMemPlugin) with session state hooks, and registered Bub tools for memory operations (mem.search, mem.save, mem.context, mem.connections, mem.timeline, mem.forget, mem.threads, mem.thread, mem.status). Changes
Sequence Diagram(s)sequenceDiagram
participant Agent as Bub Agent
participant Plugin as NowledgeMemPlugin
participant Client as NmemClient
participant CLI as nmem CLI
participant Mem as Memory System
rect rgba(100, 150, 255, 0.5)
Note over Agent,Mem: Session Load & Recall
Agent->>Plugin: load_state(session_id)
Plugin->>Client: read_working_memory()
Client->>CLI: nmem get-working-memory --json
CLI->>Mem: request working memory
Mem-->>CLI: working memory JSON
CLI-->>Client: JSON payload
Client-->>Plugin: working memory dict
Plugin->>Client: search(recall_query)
Client->>CLI: nmem search --json ...
CLI->>Mem: perform search
Mem-->>CLI: search results
CLI-->>Client: results JSON
Client-->>Plugin: recalled list
Plugin-->>Agent: state with _nmem_working_memory/_nmem_recalled
end
sequenceDiagram
participant Agent as Bub Agent
participant Tool as mem.save Tool
participant Client as NmemClient
participant CLI as nmem CLI
participant Mem as Memory System
rect rgba(100, 200, 100, 0.5)
Note over Agent,Mem: mem.save with pre-dedup
Agent->>Tool: mem.save(content, title, labels...)
Tool->>Client: search(content, labels=..., limit=1)
Client->>CLI: nmem search --json ...
CLI->>Mem: search request
Mem-->>CLI: existing memory or empty
CLI-->>Client: response
alt duplicate found
Client-->>Tool: existing memory info
Tool-->>Agent: Skipped — duplicate detected
else no duplicate
Tool->>Client: add_memory(...)
Client->>CLI: nmem add-memory --json ...
CLI->>Mem: store memory
Mem-->>CLI: new memory JSON
CLI-->>Client: response
Client-->>Tool: new memory dict
Tool-->>Agent: Saved — memory id & summary
end
end
Estimated code review effort🎯 4 (Complex) | ⏱️ ~45 minutes Possibly related PRs
Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
Bugbot Free Tier Details
You are on the Bugbot Free tier. On this plan, Bugbot will review limited PRs each billing cycle.
To receive Bugbot reviews on all of your PRs, visit the Cursor dashboard to activate Pro and start your 14-day free trial.
Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (8)
nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py (3)
125-128: Use exception chaining for JSON decode errors.♻️ Proposed fix
except json.JSONDecodeError as e: - raise NmemError(f"invalid JSON from nmem: {e}") + raise NmemError(f"invalid JSON from nmem: {e}") from e🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py` around lines 125 - 128, The JSON decode exception handler in the block that parses `raw` currently raises NmemError without chaining the original exception; update the except in that parsing code to re-raise NmemError using exception chaining (e.g., "raise NmemError(f\"invalid JSON from nmem: {e}\") from e") so the original json.JSONDecodeError is preserved for traceback and debugging while keeping the existing message.
109-113: Use exception chaining for better tracebacks.When re-raising exceptions, use
raise ... from err(orfrom Noneto suppress chaining) to preserve the exception context for debugging.♻️ Proposed fix
- except asyncio.TimeoutError: + except asyncio.TimeoutError as err: proc.kill() # type: ignore[union-attr] - raise NmemError(f"nmem timed out after {timeout}s") + raise NmemError(f"nmem timed out after {timeout}s") from err - except FileNotFoundError: - raise NmemError("nmem not found in PATH") + except FileNotFoundError as err: + raise NmemError("nmem not found in PATH") from err🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py` around lines 109 - 113, The except blocks in client.py are re-raising NmemError without exception chaining; change them to capture the original exceptions (e.g., except asyncio.TimeoutError as err and except FileNotFoundError as err), keep the proc.kill() in the timeout branch, and re-raise using "raise NmemError(f'nmem timed out after {timeout}s') from err" and "raise NmemError('nmem not found in PATH') from err" so the original traceback is preserved for debugging (refer to the exception handling in the function where proc.kill() is called and the two except blocks).
64-73: Consider addinguvxfallback fornmemcommand resolution.The current implementation only checks for
nmemin PATH. Based on learnings from other Nowledge Mem plugins, the recommended pattern includes a fallback touvx --from nmem-cli nmemwhennmemis not directly available.♻️ Proposed fix to add uvx fallback
def _resolve_cmd(self) -> str: if self._cmd is not None: return self._cmd cmd = shutil.which("nmem") if cmd: self._cmd = cmd return cmd + # Fallback: try uvx + uvx = shutil.which("uvx") + if uvx: + self._cmd = uvx + return uvx raise NmemError( - "nmem not found in PATH. Install with: pip install nmem-cli" + "nmem not found in PATH. Install with: pip install nmem-cli or uvx --from nmem-cli nmem" ) + + def _base_args(self, *, json_output: bool = True) -> list[str]: + cmd = self._resolve_cmd() + # If using uvx, prepend the package specifier + if cmd.endswith("uvx"): + args = [cmd, "--from", "nmem-cli", "nmem"] + else: + args = [cmd] + if json_output: + args.append("--json") + if self._api_url: + args.extend(["--api-url", self._api_url]) + return argsBased on learnings: "Memory backend uses
nmemCLI, with fallback touvx --from nmem-cli nmem"🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py` around lines 64 - 73, The _resolve_cmd method currently only looks for "nmem" in PATH and raises NmemError if missing; update it to add a fallback that checks for "uvx" (using shutil.which) and, if found, sets self._cmd to the composed command "uvx --from nmem-cli nmem" and returns that string; otherwise keep raising NmemError. Reference _resolve_cmd, self._cmd, NmemError and shutil.which when making the change.nowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py (3)
294-299: Consider confirming deletion with the user.The
mem.forgettool deletes a memory without confirmation. Consider adding a confirmation step or returning more context about what was deleted to prevent accidental data loss.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py` around lines 294 - 299, The mem_forget tool (function mem_forget) currently calls _get_client(context) and immediately awaits client.delete_memory(memory_id), which can cause accidental deletions; modify mem_forget to either require an explicit confirmation flag or return a structured prompt/result before deleting, e.g., accept a confirm boolean or first fetch the memory via client.get_memory(memory_id) and return its summary for user confirmation, then only call client.delete_memory(memory_id) when confirmation is present; ensure returned responses include what was deleted (memory_id and a brief summary) and keep the same tool signature/context handling with _get_client and client.delete_memory.
140-141: Consider logging dedup check failures.The silent
passin the exception handler hides potential issues. While dedup is best-effort, a debug log would help with troubleshooting.♻️ Proposed fix
except Exception: - pass # dedup is best-effort + logger.debug("dedup check failed (best-effort, continuing)")🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py` around lines 140 - 141, Replace the silent except in the dedup check in tools.py (the "except Exception: pass" block) with a debug-level log that records the exception and a short contextual message; use the module's logger (e.g., logger or logging.getLogger(__name__)) and include exc_info=True (or logger.exception at debug level) so failures during the best-effort deduplication are captured without changing behavior.
63-74: Replace EN DASH with HYPHEN-MINUS in Field descriptions.The Field descriptions contain EN DASH characters (
–) instead of standard HYPHEN-MINUS (-). While they render similarly, using ASCII hyphens is more consistent and avoids potential encoding issues.♻️ Proposed fix
- limit: int = Field(5, description="Max results (1–20).") + limit: int = Field(5, description="Max results (1-20).")Apply similar changes to lines 103, 106, and 255.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py` around lines 63 - 74, The Field descriptions in MemSearchInput (and other Field usages) use an EN DASH (–) instead of an ASCII hyphen-minus (-); update the descriptions to replace EN DASH with hyphen-minus (e.g., change "Max results (1–20)" to "Max results (1-20)") and apply the same replacement to the other Field descriptions mentioned in this file (the other occurrences noted in the review). Locate usages by the class/name identifiers like MemSearchInput and the Field(...) calls and update their description strings to use "-" everywhere.nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py (1)
136-144: Improve recall search trigger condition.The current condition
len(query.strip()) > 3is a minimal safeguard. Consider also checking that the query contains meaningful content (not just whitespace or punctuation) to avoid unnecessary API calls.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py` around lines 136 - 144, The recall trigger uses only len(query.strip()) > 3 which still fires for punctuation-only inputs; before calling self.client.search in the block around content_of(message) update the guard to ensure the query contains meaningful text (e.g., strip whitespace and punctuation and check remaining length or require at least one alphanumeric/word character) so you only call self.client.search(query[:500], limit=5) for queries with substantive content; keep the same error handling and assignment to state["_nmem_recalled"] and retain logger.debug on exception.nowledge-mem-bub-plugin/README.md (1)
29-42: Enhance tool documentation with detailed examples and response contracts.The tool table provides helpful one-line descriptions, but lacks the detailed user-facing behavior and response contract examples that would help users understand how to call each tool and what to expect in return. Consider adding a dedicated section with:
- Parameter schemas for each tool
- Example requests and responses
- Common use cases
- Error conditions and handling
For example:
### mem.search **Parameters:** - `query` (string, required): Search terms - `labels` (array, optional): Filter by labels - `since` (string, optional): Date filter (ISO 8601) **Example:**,mem.search query="API design decisions" labels=["backend"]
**Response:** Returns a list of matching memories with id, content, labels, and relevance score.Based on learnings, README.md must document user-facing behavior and response contract examples for all tools.
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/README.md` around lines 29 - 42, Update the Tools section to include a dedicated detailed documentation subsection for each tool (mem.search, mem.save, mem.context, mem.connections, mem.timeline, mem.forget, mem.threads, mem.thread, mem.status) that specifies parameter schemas (types and required/optional), example requests (call syntax) and example responses (response contract with field names and types), common use cases, and documented error conditions and handling guidance; ensure each tool entry contains at least: a Parameters list, a short Example request, an Example response showing returned fields (id, content, labels, relevance, pagination tokens, etc. where applicable), and common errors (validation, auth, not-found) so users can reliably call and parse each tool’s output.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@nowledge-mem-bub-plugin/pyproject.toml`:
- Around line 29-30: The pyproject.toml packages list currently includes
"src/bub_skills" which is not a Python package (contains only SKILL.md and no
__init__.py) and will break the wheel build; either remove "src/bub_skills" from
the packages entry in pyproject.toml, or make bub_skills a real package by
adding an __init__.py and any needed modules so it can be imported;
alternatively, keep the directory out of packages and include SKILL.md via
package-data/include-package-data configuration so the file is bundled without
declaring bub_skills as a package.
- Line 15: The dependency line currently pins a pre-release ("bub>=0.3.0a1");
change this to use the stable release (e.g., "bub>=0.2.3") unless you
specifically need the alpha features from 0.3.0a1—if the alpha is intentional,
add a short note in the project metadata or repo docs explaining why the stable
"0.2.3" cannot be used and which 0.3.0a1 features are required; look for the
dependency string "bub>=0.3.0a1" in pyproject.toml and update or document
accordingly.
---
Nitpick comments:
In `@nowledge-mem-bub-plugin/README.md`:
- Around line 29-42: Update the Tools section to include a dedicated detailed
documentation subsection for each tool (mem.search, mem.save, mem.context,
mem.connections, mem.timeline, mem.forget, mem.threads, mem.thread, mem.status)
that specifies parameter schemas (types and required/optional), example requests
(call syntax) and example responses (response contract with field names and
types), common use cases, and documented error conditions and handling guidance;
ensure each tool entry contains at least: a Parameters list, a short Example
request, an Example response showing returned fields (id, content, labels,
relevance, pagination tokens, etc. where applicable), and common errors
(validation, auth, not-found) so users can reliably call and parse each tool’s
output.
In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py`:
- Around line 125-128: The JSON decode exception handler in the block that
parses `raw` currently raises NmemError without chaining the original exception;
update the except in that parsing code to re-raise NmemError using exception
chaining (e.g., "raise NmemError(f\"invalid JSON from nmem: {e}\") from e") so
the original json.JSONDecodeError is preserved for traceback and debugging while
keeping the existing message.
- Around line 109-113: The except blocks in client.py are re-raising NmemError
without exception chaining; change them to capture the original exceptions
(e.g., except asyncio.TimeoutError as err and except FileNotFoundError as err),
keep the proc.kill() in the timeout branch, and re-raise using "raise
NmemError(f'nmem timed out after {timeout}s') from err" and "raise
NmemError('nmem not found in PATH') from err" so the original traceback is
preserved for debugging (refer to the exception handling in the function where
proc.kill() is called and the two except blocks).
- Around line 64-73: The _resolve_cmd method currently only looks for "nmem" in
PATH and raises NmemError if missing; update it to add a fallback that checks
for "uvx" (using shutil.which) and, if found, sets self._cmd to the composed
command "uvx --from nmem-cli nmem" and returns that string; otherwise keep
raising NmemError. Reference _resolve_cmd, self._cmd, NmemError and shutil.which
when making the change.
In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py`:
- Around line 136-144: The recall trigger uses only len(query.strip()) > 3 which
still fires for punctuation-only inputs; before calling self.client.search in
the block around content_of(message) update the guard to ensure the query
contains meaningful text (e.g., strip whitespace and punctuation and check
remaining length or require at least one alphanumeric/word character) so you
only call self.client.search(query[:500], limit=5) for queries with substantive
content; keep the same error handling and assignment to state["_nmem_recalled"]
and retain logger.debug on exception.
In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py`:
- Around line 294-299: The mem_forget tool (function mem_forget) currently calls
_get_client(context) and immediately awaits client.delete_memory(memory_id),
which can cause accidental deletions; modify mem_forget to either require an
explicit confirmation flag or return a structured prompt/result before deleting,
e.g., accept a confirm boolean or first fetch the memory via
client.get_memory(memory_id) and return its summary for user confirmation, then
only call client.delete_memory(memory_id) when confirmation is present; ensure
returned responses include what was deleted (memory_id and a brief summary) and
keep the same tool signature/context handling with _get_client and
client.delete_memory.
- Around line 140-141: Replace the silent except in the dedup check in tools.py
(the "except Exception: pass" block) with a debug-level log that records the
exception and a short contextual message; use the module's logger (e.g., logger
or logging.getLogger(__name__)) and include exc_info=True (or logger.exception
at debug level) so failures during the best-effort deduplication are captured
without changing behavior.
- Around line 63-74: The Field descriptions in MemSearchInput (and other Field
usages) use an EN DASH (–) instead of an ASCII hyphen-minus (-); update the
descriptions to replace EN DASH with hyphen-minus (e.g., change "Max results
(1–20)" to "Max results (1-20)") and apply the same replacement to the other
Field descriptions mentioned in this file (the other occurrences noted in the
review). Locate usages by the class/name identifiers like MemSearchInput and the
Field(...) calls and update their description strings to use "-" everywhere.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 48f593ae-6a80-4033-9c66-b5bc4550c0e2
📒 Files selected for processing (9)
nowledge-mem-bub-plugin/.gitignorenowledge-mem-bub-plugin/CHANGELOG.mdnowledge-mem-bub-plugin/README.mdnowledge-mem-bub-plugin/pyproject.tomlnowledge-mem-bub-plugin/src/bub_skills/nowledge-mem/SKILL.mdnowledge-mem-bub-plugin/src/nowledge_mem_bub/__init__.pynowledge-mem-bub-plugin/src/nowledge_mem_bub/client.pynowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.pynowledge-mem-bub-plugin/src/nowledge_mem_bub/tools.py
…ckaging, dead code - save_state: only retry create_thread on "not found" errors, not timeouts/auth; cache thread existence to skip retry on subsequent turns - bub_skills: add __init__.py so importlib can discover the skill directory - client: remove unused get_memory method - bump to v0.1.1 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
There was a problem hiding this comment.
🧹 Nitpick comments (2)
nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py (1)
166-167: SHA1 for ID generation is acceptable, but consider SHA256.Static analysis flags
sha1as insecure (S324), but this is a false positive here—SHA1 is used for deterministic ID generation fromsession_id, not for cryptographic security. The 10-char hex substring provides sufficient uniqueness for thread IDs.That said, switching to
sha256silences the linter at no performance cost and is considered modern best practice.🔇 Optional: use SHA256 to silence linter
- digest = hashlib.sha1(session_id.encode()).hexdigest()[:10] + digest = hashlib.sha256(session_id.encode()).hexdigest()[:10]🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py` around lines 166 - 167, Replace the use of hashlib.sha1 with hashlib.sha256 when deriving digest from session_id: update the digest calculation (the variable digest that currently does hashlib.sha1(session_id.encode()).hexdigest()[:10]) to use hashlib.sha256(session_id.encode()).hexdigest()[:10], leaving the thread_id construction (thread_id = f"bub-{digest}") and session_id handling unchanged so the deterministic ID behavior remains the same and the linter warning is silenced.nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py (1)
109-128: Add exception chaining for better debugging.Re-raised exceptions should chain the original cause using
raise ... from err(to preserve the traceback) orraise ... from None(to explicitly suppress it). This aids debugging when nmem failures occur.♻️ Proposed fix for exception chaining
except asyncio.TimeoutError: proc.kill() # type: ignore[union-attr] - raise NmemError(f"nmem timed out after {timeout}s") + raise NmemError(f"nmem timed out after {timeout}s") from None except FileNotFoundError: - raise NmemError("nmem not found in PATH") + raise NmemError("nmem not found in PATH") from Noneexcept json.JSONDecodeError as e: - raise NmemError(f"invalid JSON from nmem: {e}") + raise NmemError(f"invalid JSON from nmem: {e}") from e🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py` around lines 109 - 128, The NmemError raises lose the original exception context; update the exception handling in _exec and _exec_json to chain causes: in the asyncio.TimeoutError except block re-raise NmemError using "raise NmemError(f'nmem timed out after {timeout}s') from err" (or similar naming) so the timeout traceback is preserved; in the FileNotFoundError except block re-raise NmemError using "raise NmemError('nmem not found in PATH') from err"; and in _exec_json catch of json.JSONDecodeError change "raise NmemError(f'invalid JSON from nmem: {e}')" to "raise NmemError(f'invalid JSON from nmem: {e}') from e" so the original JSONDecodeError is chained. Ensure you reference the _exec and _exec_json functions and the NmemError type when making these changes.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Nitpick comments:
In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/client.py`:
- Around line 109-128: The NmemError raises lose the original exception context;
update the exception handling in _exec and _exec_json to chain causes: in the
asyncio.TimeoutError except block re-raise NmemError using "raise
NmemError(f'nmem timed out after {timeout}s') from err" (or similar naming) so
the timeout traceback is preserved; in the FileNotFoundError except block
re-raise NmemError using "raise NmemError('nmem not found in PATH') from err";
and in _exec_json catch of json.JSONDecodeError change "raise
NmemError(f'invalid JSON from nmem: {e}')" to "raise NmemError(f'invalid JSON
from nmem: {e}') from e" so the original JSONDecodeError is chained. Ensure you
reference the _exec and _exec_json functions and the NmemError type when making
these changes.
In `@nowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py`:
- Around line 166-167: Replace the use of hashlib.sha1 with hashlib.sha256 when
deriving digest from session_id: update the digest calculation (the variable
digest that currently does hashlib.sha1(session_id.encode()).hexdigest()[:10])
to use hashlib.sha256(session_id.encode()).hexdigest()[:10], leaving the
thread_id construction (thread_id = f"bub-{digest}") and session_id handling
unchanged so the deterministic ID behavior remains the same and the linter
warning is silenced.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 588ffa05-34ba-42ca-a2ad-45df5369f8be
📒 Files selected for processing (5)
nowledge-mem-bub-plugin/CHANGELOG.mdnowledge-mem-bub-plugin/pyproject.tomlnowledge-mem-bub-plugin/src/bub_skills/__init__.pynowledge-mem-bub-plugin/src/nowledge_mem_bub/client.pynowledge-mem-bub-plugin/src/nowledge_mem_bub/plugin.py
✅ Files skipped from review due to trivial changes (1)
- nowledge-mem-bub-plugin/src/bub_skills/init.py
🚧 Files skipped from review as they are similar to previous changes (2)
- nowledge-mem-bub-plugin/pyproject.toml
- nowledge-mem-bub-plugin/CHANGELOG.md
Member
Author
|
Let's merge it and working on new PR for iterations! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Native Bub integration via nowledge-mem-bub package:
chore: remove .pyc files and add .gitignore for Bub plugin
Note
Medium Risk
Introduces a new Bub plugin that shells out to the
nmemCLI and can automatically ingest session turns into external storage, so failures or misconfiguration could affect runtime behavior and data capture/privacy expectations.Overview
Adds an initial
nowledge-mem-bubBub plugin package that connects Bub sessions to Nowledge Mem via thenmemCLI, including packaging (pyproject.toml), docs (README.md,CHANGELOG.md), and a bundlednowledge-memskill.Implements Bub hooks (
system_prompt,load_state,save_state) to provide agent guidance, optionally auto-inject Working Memory + recalled memories per turn (NMEM_SESSION_CONTEXT), and incrementally capture each user/assistant turn into a Nowledge Mem thread (NMEM_SESSION_DIGEST).Registers 9 new
mem.*tools (mem.search,mem.savewith best-effort dedupe,mem.context,mem.connections,mem.timeline,mem.forget,mem.threads,mem.thread,mem.status) backed by an asyncNmemClientwrapper supporting config/env-based remote access.Written by Cursor Bugbot for commit 23a371e. This will update automatically on new commits. Configure here.
Summary by CodeRabbit
New Features
Documentation
Chores