llms.txt for agents

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

With readOnly, openWorld, and idempotent annotations already covering safety and side effects, the description adds valuable behavioral context: the HEAD-probed nature of the llms-full.txt check, the fact that every attempted rung is returned, and that misses are diagnosable. This goes well beyond what annotations alone communicate.

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 sentences with no filler. The core behavior and output contents are front-loaded, and the usage guidance follows naturally. Every clause earns its place, and the structure gives the agent the key decision information early.

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?

Despite lacking an output schema, the description conveys what will be returned (title, summary, H2 links, Optional section, llms-full.txt indicator) and how failures behave ('every rung tried is returned so a miss is diagnosable'). Combined with the annotations and the sibling-routing guidance, an agent has enough context to invoke it correctly.

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 the schema already documents all three parameters. The description reinforces the URL-centric purpose and mentions the parsed output, but it does not add meaningful detail about the parameters themselves beyond what the schema provides. Baseline 3 is appropriate.

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 uses a specific verb ('Find and parse') with a clear resource ('the llms.txt that covers a site or page URL') and enumerates the exact parsed contents (title, summary, H2 sections, Optional section, llms-full.txt existence). It also differentiates from the sibling llms_page by saying 'Use first, before llms_page,' so an agent can distinguish it without opening the schema.

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 explicitly states when to use this tool: 'Use first, before llms_page, whenever the user needs a product or library's own docs.' This provides clear routing relative to llms_page. It does not explicitly mention when llms_search would be preferable, so the guidance is strong but not fully exhaustive.

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

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

The description thoroughly discloses the internal fallback chain (Accept negotiation, alternate markdown mirror, file conventions, final HTML-to-markdown conversion) and the bounded pagination behavior. This goes well beyond the readOnly and idempotent annotations, giving the agent a realistic expectation of how the tool behaves and what it may return.

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 a single, information-dense sentence that front-loads the core purpose and pagination behavior before the usage note. Every clause adds value, and there is no filler or repetition of schema details.

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 tool is moderately complex, but the description covers its purpose, input source, fallback behaviors, pagination, and the role of task_context. No output schema exists, yet the mention of 'truncated result' in the offset parameter plus the markdown return behavior is sufficient 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 coverage is 100%, so the parameters are already fully documented in the input schema. The description adds helpful context about pagination and the source of URLs, but it does not add substantial new parameter-level meaning beyond what the schema already provides.

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 action ('Return one docs page as markdown') and a clear resource ('docs page'), with pagination behavior explicitly mentioned. It also distinguishes itself from siblings by referencing llms_index as the source of URLs, making its role in the tool family clear.

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 explicit guidance: 'Use on the URLs llms_index returns, or on any docs URL you already hold.' This clearly states when to use the tool. It does not explicitly mention llms_search as an alternative or state when not to use it, but the context is still clear enough for an agent.

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

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

Annotations already convey read-only, open-world, and idempotent behavior. The description adds meaningful context beyond those annotations: the weighted fields searched (headings ×3, titles ×2, notes ×1), the per-site scoping, and the fact that a site may be indexed on demand. No contradiction with the annotations is evident.

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 sentences deliver purpose, scope, filtering behavior, and a usage recommendation with no filler. The most important information is front-loaded, and the parenthetical weight details are compact and useful.

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 search tool with no output schema, the description communicates the essential returned content: matching docs links and their covering index. It does not describe result ordering, ranking, or exact response shape, but the limit and max_bytes parameters plus the stated output are enough for an agent to call it correctly.

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?

With 80% schema description coverage, the schema already explains most parameters. The description adds extra meaning for 'query' by revealing the weighted index fields and for 'site' by clarifying single-site scoping with on-demand indexing, which is not fully captured 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 states a specific action ('Search the llms.txt indexes') and a concrete result ('get the matching docs links with their covering index'). It also distinguishes itself from reading the whole index, which maps to the sibling llms_index tool.

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?

It explicitly says when to use this tool: 'jump straight to the right page of a docs set instead of reading the whole index,' clearly positioning it against the index-reading sibling. It also explains the 'site' restriction and on-demand indexing, though it does not explicitly spell out when to prefer llms_page.

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