Preprocesses loaded ARF matrices by applying imputation, normalization, QC filtering, and drift correction, then updates the session and returns a report for downstream PCA or differential analysis.
ロード済み ARF 行列に前処理レシピを適用し、session を更新して報告を返す。
以降の PCA/差次的解析は session_state.session.feature_matrix（前処理後）を消費する。
| Name | Required | Description | Default |
|---|---|---|---|
| props | No | ||
| impute | No | half_min | |
| normalize | No | none | |
| max_qc_rsd | No | ||
| drift_correct | No | ||
| blank_min_fold | No |
Changes observed during successful MCP inspections.
v0.1.0Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description does disclose valuable behavior beyond the schema: it mutates the session and writes the result to session_state.session.feature_matrix, which downstream steps read. However, that stated session update sits in tension with the annotation readOnlyHint=true, so the agent receives conflicting signals about side effects. It also says a 'report' is returned without describing its content.
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?
Two short sentences with no filler, and the core action is front-loaded before the downstream-dependency note. It is efficiently written, though the terseness is partly a symptom of the missing parameter and return-value detail.
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 six-parameter preprocessing tool with no output schema and no documented parameters, the description covers only the high-level purpose and the session dependency. It does not explain the recipe options, defaults, valid ranges, or what the returned report contains, leaving most of what an agent needs to call it correctly undocumented.
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?
Six parameters exist with 0% schema description coverage and no enums, and the description mentions none of them – not props, impute, normalize, max_qc_rsd, drift_correct, or blank_min_fold. With the schema carrying no semantics and the description adding nothing, an agent cannot tell what a 'recipe' consists of or what values are legal.
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 states a specific verb and resource – applying a preprocessing recipe to an already-loaded ARF matrix – and adds the concrete effect of updating the session. It is distinguishable from siblings like arf_pca_preprocessed or dataset_preprocess by naming the preprocessing step itself, though it never explicitly differentiates itself from those near-neighbours.
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?
Usage is only implied: the second sentence says downstream PCA/differential analyses consume session_state.session.feature_matrix after preprocessing, which hints this runs before those steps. There is no explicit 'use this when...', no statement of prerequisites (which matrix must already be loaded), and no named alternative such as dataset_preprocess.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.