jitprompt gives your agents just-in-time instructions. Upload your docs as they are, and your agent gets the right section over MCP at the moment it needs it, instead of 20k tokens of system prompt every turn.
Scored across 3 tools
get_toc (browse the map), search_instructions (find sections by topic), and get_section (read a section in full) each have a distinct role, and the descriptions explicitly cross-reference each other to steer selection. There is mild overlap because get_toc can also take a query or document_id to list sections, but the docs push users to search_instructions for lookups, keeping boundaries clear.
All three tools use a clean snake_case verb_noun pattern: get_section, get_toc, search_instructions. Two 'get_' readers plus one 'search_' finder is a predictable, readable convention with no mixing of styles.
Three tools is a tight, well-scoped set for a read-only instruction-retrieval surface: map/navigate, search, and fetch. Each earns its place with no redundant or filler tools.
The set covers the full retrieval lifecycle for its domain: discovering documents/sections (get_toc), locating relevant ones (search_instructions), and reading them (get_section). Since instructions are authored elsewhere, no write operations are expected; only a batch/multi-section fetch is absent, which agents can work around with repeated calls.
Fetch the full text of one instruction section by its id (from get_toc or search_instructions), with the document it belongs to. Always read the whole section before acting on it — rules and their exceptions live together.
| Name | Required | Description | Default |
|---|---|---|---|
| section_id | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the behavioral burden. It discloses a useful trait beyond the schema (the parent document is returned alongside the section) and implies a read-only fetch, but says nothing about invalid-id handling, truncation, or size limits for a 'full text' payload.
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, no filler. The identity/provenance statement comes first, followed by the actionable read-the-whole-section caution, which is exactly the kind of front-loading an agent needs.
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?
An output schema exists, so return values need no explanation; the tool has one required parameter whose origin is documented, and the behavioral caveat about rules and exceptions is included. Nothing essential to correct invocation is missing.
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 0% and the single parameter has no description, so the description must compensate — and it does, explaining that section_id is an instruction-section id sourced from get_toc or search_instructions. It doesn't give format details (e.g., id shape), which keeps this short of a 5.
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?
States a specific verb (fetch) and resource (full text of one instruction section), plus what else is returned (the parent document). It also disambiguates from siblings by naming get_toc and search_instructions as the source of valid ids.
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?
Gives clear workflow guidance — obtain the id from get_toc or search_instructions, then read the whole section before acting because rules and exceptions live together. There is no explicit when-not guidance, but the usage context is unambiguous.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Map of the instruction documents available to you: one entry per document with a one-line summary of what it covers and how to use it. Small documents list their sections inline (id, title, summary); large ones show a section count and how to expand them. Call this at the start of a conversation and again whenever the conversation moves to a topic or system you haven't checked guidance for. Cheap; safe to call repeatedly.
In a large collection the map is condensed (document lines, titles, or just an overview) and says so; then pass query.
query: what you are about to do (e.g. "checking a product's eligibility rules before an application"). Lists only the documents relevant to it, best first — the map for the task at hand.
document_id: pass a document's id (as shown in the map) to list every section of that one document. Do this to browse a document you need to read through. To find a specific rule, identifier, error code or table, call search_instructions instead — it is cheaper than expanding a large document and reading its section list.
Ids in the output are short prefixes; pass them as-is to get_section or get_toc.
| Name | Required | Description | Default |
|---|---|---|---|
| query | No | ||
| document_id | No |
| Name | Required | Description |
|---|---|---|
| result | Yes |
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 advertises cost ('Cheap; safe to call repeatedly'), describes output variability (small docs inline vs large docs condensed, and that the output 'says so'), and explains how to widen results. It stops short of mentioning auth/permission requirements or rate limits, but covers the operationally relevant traits.
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?
Purpose and usage are front-loaded before parameter guidance, and each paragraph earns its place with concrete, actionable detail. It is dense and prose-heavy rather than scannable, but there is little true 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?
An output schema exists, yet the description still orients the agent on the shape of the result (entries, inline sections vs counts, condensed collections) and covers both parameters plus sibling routing. Nothing needed to invoke this correctly is missing.
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 0% and both params are undescribed in the schema, so the description must compensate and it does: query is explained as 'what you are about to do' with a concrete example and its return semantics ('best first'), and document_id is explained as listing every section of one document with a pointer to search_instructions for narrow lookups. It also clarifies the id format (short prefixes, pass as-is).
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?
States a specific resource ('map of the instruction documents available to you') and exactly what each entry contains (one-line summary, section lists or counts). It explicitly distinguishes itself from search_instructions and get_section by naming what each of those is for, so an agent can route correctly without opening any 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?
Gives explicit when-to-call ('at the start of a conversation and again whenever the conversation moves to a topic... you haven't checked guidance for') and a clear when-not path ('To find a specific rule... call search_instructions instead — it is cheaper'). The document_id branch names its own alternative behavior for browsing a whole document.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Find instruction sections relevant to what you are ABOUT TO DO. Call this before EVERY new topic, system, or data source you engage with — not just once per conversation. If the user's question shifts to a different system, product or task than the one you last checked guidance for (e.g. from refunds to account closures, or from one product to another), call this again before acting, even if you already called it earlier in this conversation. Describe your plan and the guidance you need — e.g. 'writing a refund email to an enterprise customer' or 'checking the eligibility rules for a product application' — NOT the user's question verbatim. Name the product, system or identifier when you know it. Returns whole sections grouped by document, best document first; a small document comes back whole, so read it as one piece.
| Name | Required | Description | Default |
|---|---|---|---|
| plan | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the full burden, and it does disclose result behavior: sections are grouped by document with 'best document first,' and small documents come back whole. It does not mention result limits, pagination, or how broad a 'plan' match can be, which is the remaining gap for a search tool.
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 core instruction is front-loaded and every sentence is actionable. The re-emphasis that you must call again 'even if you already called it earlier in this conversation' is slightly repetitive, though arguably deliberate reinforcement for a tool that must not be called once-per-conversation.
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?
An output schema exists, so return values need not be explained, and the description still adds ranking context. With one required parameter, no annotations, and no nesting, the description covers invocation timing, input construction, and result shape — nothing needed to call it correctly is missing.
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 0% for the single 'plan' parameter, and the description fully compensates: it explains what to write (your plan plus the guidance you need), gives three concrete example strings, warns against pasting the user's question verbatim, and instructs naming the product/system/identifier.
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 specific verb+resource ('Find instruction sections') and immediately narrows scope with 'relevant to what you are ABOUT TO DO.' The search-by-relevance framing is inherently distinguishable from the get-by-id / get-structure siblings get_section and get_toc.
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 is explicit about when to call ('before EVERY new topic, system, or data source'), when to re-call ('if the user's question shifts to a different system, product or task'), and gives concrete transition examples (refunds to account closures). This is textbook when/when-again guidance for a retrieval tool.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
get_sectionget_tocsearch_instructions
Add one secure layer between your agents and this server.