$ loading_
Give AI coding agents searchable local project memory with safe structured updates.
This MCP tool is described as local project memory with no required secrets and no declared remote endpoints. Given its open-source Apache 2.0 status, overall risk appears relatively low, but as executable third-party code with weak evidence of adoption and maintenance, its local data access and actual implementation still warrant caution.
The material explicitly states that no keys or environment variables are required. No API tokens, cloud credentials, or account authorization are requested, so credential exposure risk appears low.
The material declares no remote endpoints, and the description emphasizes local project memory. There is no stated evidence of sending user data to external services.
The system flags it as executable code in an MCP tool; running such a tool typically involves local process execution. The material does not specify its exact system-call scope, which is a normal caution for this class of tool, with no additional high-risk red flags shown.
The description says it provides 'local, searchable project memory' with Markdown as the source of truth, implying it likely reads and may write project memory files. This aligns with its stated function, but the missing README leaves directory boundaries, file patterns, and workspace-only restrictions unverified.
There is a public GitHub repository under Apache 2.0, making the source in principle auditable, which is a meaningful risk reducer. However, it comes from a third-party registry, shows 0 stars, and has unknown maintenance status, so trust and ongoing maintenance evidence are limited and the repo and dependencies should be independently reviewed.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "agent-memory" yet — see the docs or source repo.
Search agent-memory for this project's coding standards, folder conventions, and commit message format, then summarize them as bullet points.
A concise summary of existing development conventions to align work before coding.
Write today's decision to switch the auth module to JWT with 30-day refresh tokens into agent-memory using a structured update, including rationale, scope, and follow-up tasks.
A properly structured project decision entry that can be searched and referenced later.
Before implementing the payment webhook feature, extract related module background, past issues, API constraints, and open tasks from agent-memory to create a development context brief.
A task-oriented context brief that reduces repeated investigation before implementation.
Give AI coding agents persistent semantic memory with embedding-based retrieval.
Give local AI coding assistants persistent memory and semantic code search.
Store and retrieve project memory for coding assistants via MCP tools.
Give coding agents persistent, evidence-backed project memory across sessions.
Store coding history, decisions, and fixes for AI agents in your repo.
Manage durable project memory in AGENTS.md for AI agents.