mc_kit_memory

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

With no annotations, the description carries the full disclosure burden and does well: it states that data survives sessions and restarts, that `README.md` is loaded into server instructions (an important side effect), that delete on a directory path removes the whole directory, and that read on an image returns the image itself. The remaining gap is return-format detail beyond the outline of search hits and index contents, and any locking or concurrency behavior.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

The purpose and storage layout are front-loaded, followed by a scannable bulleted action list, and then path syntax and the no-shell caveat. Almost every line carries information, although the bullets sit adjacent to the already self-documenting schema descriptions, which creates some redundancy.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 16-parameter, annotation-free, output-schema-free tool, the description covers routing all nine actions, the on-disk layout, the README index side effect, path/topic/server resolution, and the safety boundary. It does not describe return shapes in much detail, but the variants for index, kinds, and search are outlined enough for correct invocation.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100%, so all 16 parameters are already documented, including their per-action applicability (record/append, put, search) and the required-fields rules. The description adds only the interaction rule that you supply `path` or `topic`, with `server` defaulting to the current server and `_global/` used when omitted. This is useful but largely restates what the schema already says about `topic`, `path`, and `server`.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

The description opens with a concrete verb and resource combination: 长期记忆（跨会话、重启后还在）, stored under `.mcp-mc-agent/` with server-indexed subfolders. It then enumerates all nine actions with a one-line statement of what each does, for example index 看总览, delete 删文件, and search 跨文本文件搜关键词. No sibling tool deals with persistence, so an agent can identify this tool's role immediately.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description gives strong selection guidance among the nine actions: record is marked 推荐 and mentions its per-key accumulation semantics, append is explicitly text-file-only, write is for restructuring, and put for copying workspace files. The warning that the tool 只读写文件，不执行任何东西 sets a hard boundary on what it may be used for. It does not say when memory should or should not be consulted in the agent's workflow, but the action-level routing is unusually explicit.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.