Score the privilege a Chrome MV3 extension takes from its manifest, and diff permission sets.
Scored across 3 tools
Each tool has a clearly distinct purpose: analyzing a full manifest, comparing two versions, and explaining a single permission. There is no overlap, so an agent can easily select the right tool.
All tool names follow the exact verb_noun pattern: analyze_manifest, compare_permission_sets, explain_permission. The naming is uniform and predictable.
With 3 tools, the server is well-scoped for its purpose. Each tool provides a distinct and necessary function, and the count sits comfortably within the ideal range.
The set covers the core workflows of permission-risk analysis: full manifest evaluation, version comparison, and individual permission explanation. There are no obvious dead ends or missing operations for the stated domain.
Static privilege analysis of a Chrome MV3 manifest.json. Returns a 0-100 risk score, the permissions and host patterns that drive it, dangerous permission combinations, and MV3 policy problems (remote code, unsafe-eval, web_accessible_resources). Content-script matches are counted as host access even when host_permissions is empty.
| Name | Required | Description | Default |
|---|---|---|---|
| manifest | Yes | The manifest.json content, as a JSON object or a JSON string. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden of behavioral disclosure. It explains the tool is static (read-only), lists the key outputs, and adds a nuanced behavioral detail: content-script matches count as host access even when host_permissions is empty. This goes beyond a simple purpose statement, though it does not cover edge cases like invalid manifest handling.
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 long, front-loaded with the core purpose, and provides essential details without fluff. The second sentence adds an important nuance, making every word earn its place. No unnecessary or repetitive content.
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?
Without an output schema, the description enumerates the primary outputs (risk score, permissions, host patterns, combinations, policy problems), which is sufficient to set expectations. It also notes a special counting behavior for content scripts. It does not discuss error handling or the exact response structure, but for a static analysis tool with a single parameter, it covers the essential context well.
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 provides 100% coverage with a clear description of the 'manifest' parameter as JSON object or string. The tool description adds further semantic meaning by specifying the expected input is a Chrome MV3 manifest, which imposes domain constraints on the parameter value. This raises the value above the schema-alone baseline.
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 states the tool performs static privilege analysis of a Chrome MV3 manifest.json and enumerates its outputs (risk score, permissions, host patterns, dangerous combinations, MV3 policy problems). This is specific enough to distinguish it from siblings like compare_permission_sets and explain_permission, which serve different purposes.
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 clearly implies when to use this tool—when analyzing a Chrome MV3 manifest for privilege-related risks. It provides strong contextual cues and does not exclude any specific scenarios, but it does not explicitly mention alternatives or state 'use this instead of X'. Thus it has clear context without explicit exclusions.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Compares the permissions and host patterns of two versions of an extension. Reports the score delta, what was added or removed, and whether the change widens the install-time warning set, which makes Chrome disable the extension for existing users until they re-accept.
| Name | Required | Description | Default |
|---|---|---|---|
| after | Yes | API permissions in the new version. | |
| before | Yes | API permissions in the current published version. | |
| after_hosts | No | host_permissions in the new version. | |
| before_hosts | No | host_permissions in the current published version. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations, the description carries the burden of disclosure. It reveals key behavioral traits: it reports added/removed permissions and whether the change widens the install-time warning set, which can cause Chrome to disable the extension for existing users. This is a non-obvious consequence that goes beyond the basic 'compares' verb, though it stops short of detailing whether the tool itself has side effects or requires specific access.
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, front-loaded with the main action and followed by precise output details. No redundant or filler text; every sentence earns its place.
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?
Given the tool's moderate complexity and no output schema, the description adequately summarizes the return values (score delta, additions/removals, warning-set change). It also covers the practical consequence of the comparison, making it self-sufficient for an agent to understand what the tool does and what it returns.
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 input schema has 100% parameter description coverage, so the schema already explains each parameter (before, after, before_hosts, after_hosts). The description adds no new parameter-level meaning beyond what the schema provides, so the baseline score of 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 clearly states the tool compares permissions and host patterns of two versions of an extension, which is a specific verb+resource combination. It also distinguishes from siblings (analyze_manifest, explain_permission) by focusing on comparison between versions and reporting deltas.
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 implies usage for comparing two versions of an extension, and the outputs (score delta, added/removed) make the context clear. It does not explicitly mention alternative tools or when not to use, but the purpose is self-evident and differentiates from siblings.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Returns the privilege weight (0-10) for a single Chrome extension permission or host pattern, what it actually grants, whether it triggers an install-time warning, and the narrower alternative if one exists.
| Name | Required | Description | Default |
|---|---|---|---|
| permission | Yes | A permission name such as cookies, or a host pattern such as <all_urls>. |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
With no annotations provided, the description carries the full burden of behavioral disclosure. It transparently explains what the tool returns, including weight, grants, warning status, and alternatives. It does not mention edge cases like invalid permission names or host patterns, but for a read-only lookup tool this is a strong, non-misleading description.
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, front-loaded sentence that starts with the core verb and lists all key output aspects in a structured, readable way. Every element earns its place, with no fluff or repetition.
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?
This is a simple, single-parameter tool with no output schema, and the description fully enumerates the kinds of information returned. It is sufficiently complete for an agent to select and invoke the tool 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 input schema already fully describes the single 'permission' parameter, including examples of both permission names and host patterns. The description reinforces the same idea ('permission or host pattern') but adds no parameter-specific detail beyond what the schema already covers, so a baseline of 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 ('Returns') and clearly identifies the resource ('a single Chrome extension permission or host pattern'). It enumerates the exact information provided (privilege weight, what it grants, install-time warning, narrower alternative), which fully distinguishes it from sibling tools like analyze_manifest and compare_permission_sets.
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 clearly implies this tool is for explaining one permission and lists what it returns, giving the user context for when to invoke it. However, it does not explicitly mention when not to use it or refer to alternative tools such as analyze_manifest or compare_permission_sets.
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.
analyze_manifestcompare_permission_setsexplain_permissionSigned security scores for what an AI agent runs and reads: skills, MCP servers, prompts, tokens.
Threat intel + your scans/findings/Shield posture. CVE, EPSS, KEV, package vuln lookup, DAST.
Risk-scan a diff, flag AI-generated-code tells, find secrets. 5 of 7 tools need no account.
Supply chain risk scoring for npm, PyPI, Cargo, and Go. 9 tools. Behavioral signals.
Add one secure layer between your agents and this server.
