Detect prompt injection attempts in text using guardrail patterns. Also retrieves live guardrails_blocked stats from the gateway.
| Name | Required | Description | Default |
|---|---|---|---|
| text | Yes | Text to analyze for injection attempts |
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Annotations already declare readOnlyHint=true and idempotentHint=true, so the safety profile is known. The description adds value by disclosing an unexpected side behavior—retrieving live guardrails_blocked stats—and briefly mentions the method ('guardrail patterns'). This goes beyond structured fields, though it omits details about how stats are delivered or whether they affect the detection output.
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 two sentences with no filler. The first sets the primary action, and the second discloses a relevant secondary behavior. Both earn their place, and the structure is front-loaded with the core purpose.
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 simple one-parameter tool with an output schema and read-only annotations, this description covers the purpose, primary input, and a notable side effect. The only minor gap is the lack of elaboration on how the retrieved stats are presented, but the output schema likely handles that. Overall it is complete enough for an agent to invoke 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?
The schema documents the single 'text' parameter with 100% coverage, so the description adds no new semantics beyond restating the parameter's purpose. It does not clarify format, length limits, or encoding, but the baseline of 3 is appropriate when the schema is sufficient.
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 clearly identifies the core action—detecting prompt injection attempts in text—using a specific verb and resource. However, the second sentence introduces a second purpose (retrieving live guardrails_blocked stats), which slightly dilutes focus and could confuse the agent about the tool's primary function. It still distinguishes itself from siblings by naming 'injection' specifically.
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?
There is no explicit guidance on when to use this tool versus alternatives like thinkneo_evaluate_guardrail or thinkneo_check. No exclusions, prerequisites, or example scenarios are provided, so the agent must infer applicability purely from the name and vague phrasing.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Add one secure layer between your agents and this server.
Many tools have overlapping purposes, such as the multiple safety-check tools (thinkneo_check, thinkneo_detect_injection, thinkneo_evaluate_guardrail) and the many cost/reporting tools (thinkneo_agent_roi, thinkneo_decision_cost, thinkneo_business_impact). An agent would struggle to reliably pick the correct tool for a given intent. The boundaries between dashboard, audit, and reporting tools are particularly fuzzy.
All tools share the thinkneo_ prefix, but the remaining naming is inconsistent: some follow verb_noun (check_spend, list_alerts), some use noun_verb (compliance_generate, alert_rule_create), and others are bare nouns (business_impact, cache_status). This mix makes it hard to predict tool names based on action and object.
With 68 tools, this is an extremely large surface area, far beyond the typical 3-15 well-scoped set and even beyond the 25+ heavy threshold. Even for a comprehensive enterprise platform, the sheer number overwhelms an agent's ability to choose effectively. It feels like a kitchen-sink approach rather than a curated toolkit.
The toolset covers a wide range of governance, observability, and cost-management features, but there are notable lifecycle gaps: SLAs can be defined but not updated or deleted, alert rules lack an update operation, and registry entries have no remove/unpublish. Also, policy management is limited to checking, with no create/update tool. The memory tools feel out of place and lack a delete operation.
