Server Quality Checklist

Tool Scores

• list_modelsA
  4.3/5.0
  Behavior4/5

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

  Annotations cover safety (readOnly, idempotent, non-destructive), but the description adds valuable behavioral context about what state is accessed ('bridge state', 'agent info') and what the function returns. It complements the annotations by explaining the semantic content of the operation.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  Two efficient sentences with zero waste. First sentence establishes the operation and scope; second sentence details return contents. Information is front-loaded and appropriately sized for a parameterless introspection tool.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a zero-parameter, read-only listing tool with an output schema present, the description is complete. It explains what resource is queried and what data categories are returned, which is sufficient given the low complexity and existing output schema.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  The input schema has zero parameters, establishing a baseline of 4 per the rubric. The description appropriately doesn't mention parameters since none exist, maintaining focus on the operation's purpose and return values.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description uses a specific verb (List) with clear resources (Gemini models, bridge state). It distinguishes itself from sibling 'list_sessions' by explicitly naming 'Gemini models' and 'bridge state' as targets, and from 'gemini' (likely a generation tool) by being read-only/listing focused.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  While the description implies usage by explaining what is returned (known models, active model, agent info), it lacks explicit when-to-use guidance or comparison to siblings. It doesn't state, for example, to use this before calling 'gemini' to check available models.

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

• list_sessionsA
  4.3/5.0
  Behavior4/5

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

  Adds valuable return-value context beyond annotations: specifically enumerates returned fields (workspace path, session ID, turn count, model). Also clarifies scope ('managed by the bridge') complementing openWorldHint=false. Does not contradict safety annotations.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  Two sentences, zero waste. First sentence establishes operation, second previews return payload. Information density is optimal with no redundancy.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  Fully adequate for a zero-parameter listing tool. Preview of output fields (despite existence of output schema) provides helpful context without being redundant. Annotations cover behavioral safety profile.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Zero parameters present, meeting the baseline score of 4. The description correctly avoids parameter discussion since none exist.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  Excellent specificity: 'List' (verb) + 'active ACP sessions' (resource) + 'managed by the bridge' (scope). Implicitly distinguishes from sibling 'reset_session' (list vs modify) and 'list_models' (sessions vs models).

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines3/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  States what it does clearly, providing implied usage context, but lacks explicit when-to-use guidance or contrasts with 'reset_session' (e.g., 'use this before resetting to verify session ID').

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

• reset_sessionA
  4.5/5.0
  Behavior4/5

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

  Annotations declare destructive=true, idempotent=true, readOnly=false. Description valuably adds the session lifecycle context (clearing now creates fresh session on next gemini call) and explains the workspace-scoping vs global-scope behavior, which helps the agent understand the tool's side effects beyond the safety flags.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness5/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  Three sentences with zero waste: (1) Action definition, (2) Behavioral consequence linking to sibling, (3) Parameter instruction. Front-loaded with the core operation. Every sentence earns its place by conveying distinct essential information.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  Appropriately complete for a low-complexity (1 param) destructive operation with 100% schema coverage, existing output schema, and rich annotations. Description covers scope, side effects, and parameter usage without needing to replicate return value documentation.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  Schema coverage is 100%, so fields are technically documented. Description adds semantic framing: 'or omit to reset all sessions' clarifies the empty string default behavior described in the schema, providing usage-oriented guidance that helps the agent understand the parameter's optional nature and global fallback behavior.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  States specific action 'Reset (clear)' with explicit resource 'ACP session' and scope 'workspace'. Critically references sibling tool 'gemini' ('The next gemini call...'), clarifying the tool's relationship to the session lifecycle and distinguishing it from list_sessions.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines4/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  Provides clear usage context by explaining the deferred effect on subsequent 'gemini' calls and parameter behavior ('Pass workspace path, or omit to reset all sessions'). Lacks explicit 'when not to use' guidance or named alternatives, but the causal explanation implicitly guides selection.

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

• geminiA
  4.6/5.0
  Behavior4/5

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

  Annotations declare destructive/open-world capabilities; the description adds critical runtime context missing from annotations: automatic session reuse with turn-count eviction, automatic fallback to flash-preview on rate limits, and detailed explanation of return fields (thought, tool_calls, plan) that describe the agent's internal reasoning. It could clarify the scope of possible destruction (file edits vs API calls) more explicitly.

  Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

  Conciseness4/5

  Is the description appropriately sized, front-loaded, and free of redundancy?

  The description is well-structured with clear visual hierarchy: single-line purpose statement, bulleted return structure, and bulleted best practices. Every section serves a distinct purpose. Minor deduction for the return structure list being somewhat lengthy, though justified by the complex nested return type of an agent invocation.

  Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

  Completeness5/5

  Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

  For a complex 7-parameter agent-invocation tool with destructive/open-world annotations, the description is comprehensive. It covers input parameters, return value semantics (7 distinct fields), session lifecycle, error handling strategies, and integration patterns (MCP server filtering, image ContentBlocks). Sufficient given the tool's complexity and existing schema richness.

  Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

  Parameters4/5

  Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

  With 100% schema coverage, the baseline is 3. The description adds significant value through the Best Practices section, which provides semantic guidance on model selection (complex vs simple tasks) and approval_mode implications (yolo meaning 'auto-approve all') that raw schema descriptions don't convey. It effectively guides the agent toward correct parameter combinations.

  Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

  Purpose5/5

  Does the description clearly state what the tool does and how it differs from similar tools?

  The description opens with a precise action statement ('Invokes Gemini via ACP') specifying the protocol, resource, and task type ('AI-driven tasks'). It clearly distinguishes from siblings like list_models or reset_session by being the primary execution tool versus management utilities.

  Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

  Usage Guidelines5/5

  Does the description explain when to use this tool, when not to, or what alternatives exist?

  The 'Best practices' section provides explicit guidance: model selection criteria (pro for complex, flash for simple), approval_mode behavior (yolo vs auto_edit vs default), automatic retry logic on 429 errors, and specific parameter usage patterns (image_path for vision, context for resources). It names concrete alternatives (flash model for retries) and safety thresholds.

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