memoryguard_build_and_enrich

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

With no annotations, the description carries the full burden of disclosing behavior. It does reveal an important behavioral trait: in host mode the calling LLM itself must perform classification and translation, and the tool may require a recursive call after apply_enrichments. It also warns 'Do not require a separate AI-整理 button', which is a useful anti-footgun note. However, it does not disclose whether the tool persists data, whether it is destructive, what side effects occur, or what the return value looks like — significant gaps in the absence of annotations.

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 description is four short sentences, front-loaded with the core action ('Build memory projection') followed by the default mode and conditional workflow. Every sentence adds operational value: default behavior, the enrichment loop, the CLI variant, and a UX guardrail. The only minor issue is the terse 'AI-整理' jargon, but there is no wasted wording.

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?

The description provides a usable workflow but leaves clear gaps for a complex tool with 7 parameters, no annotations, and no output schema. It does not define what a memory projection is, does not explain the reconstructed vs native distinction in the 'mode' parameter, and does not describe the expected return value. The condition 'pending_tasks / host_action_required' is mentioned but not specified, and the recursive call is only partially explained. Adequate as a minimum viable description, but not complete.

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 coverage is 100%, so the baseline is 3. The description adds meaningful semantics beyond the schema: it explains that enrich_mode=host means 'YOU are the LLM' and enrich_mode=cli uses a chosen Agent CLI, giving operational meaning to llm_cli and llm_agent. It also clarifies the default enrich_mode. It does not add value for mode, workspace, share_group_id, or agent_instance_id, but those are already described in the schema.

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 'Build memory projection', a specific verb + resource, and the workflow ('call apply_enrichments, then call this again') positions it as the orchestration entry point, distinguishing it from sibling enrichment and projection tools. However, it never explains what a 'memory projection' is, and it does not explicitly differentiate itself from closely related siblings like memoryguard_apply_enrichments or memoryguard_projection_status.

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 conditional usage guidance: 'Default enrich_mode=host: YOU are the LLM' and 'Multi-agent GUI may pass enrich_mode=cli with a chosen Agent CLI', which tells the agent when host vs cli mode is appropriate. It also provides explicit chaining instructions — 'If pending_tasks / host_action_required, immediately classify+translate, call apply_enrichments, then call this again' — which is practical when-to-invoke guidance. It does not list exclusions or compare against sibling alternatives, but the context is clear enough.

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