Server Quality Checklist

Tool Scores

• codex_read_threadA
  4.4/5.0
  Behavior5/5

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

  Beyond the readOnly and idempotent annotations, the description discloses that it reads directly from local JSONL files, needs no API call, consumes no quota, and works even when rate-limited. This adds significant behavioral context about its underlying mechanism and operational characteristics.

  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?

  The description is concise and front-loaded with the primary purpose in the first sentence. It includes only relevant details about the local file reading, quota-free operation, and default behavior, with no redundant or filler content.

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

  Completeness4/5

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

  The description is fairly complete for a read-only tool with annotations and full schema coverage. It explains the key use case, mechanism, and default behavior. However, it lacks details about the return format (e.g., JSON structure) or error handling, which would be useful given there is no output schema.

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

  Parameters3/5

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

  The input schema has 100% parameter coverage with detailed descriptions for limit, max_chars, thread_id, and include_reasoning. The description only adds that it returns the most recent messages by default, which is already implied by the schema's 'default: 50' and 'counting back from the most recent'. No additional parameter meaning is provided beyond the schema.

  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 clearly states the tool's function: 'Read the message history of any Codex thread'. It identifies the specific resource (Codex thread), the action (read), and distinguishes from siblings by covering both interactive and programmatic threads. It also mentions reading from ~/.codex/sessions, which is a unique characteristic.

  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?

  The description provides clear context on when to use the tool: it can read any thread, requires no API call, consumes no quota, and works under rate limits. However, it doesn't explicitly mention alternatives like codex_local_sessions for listing threads or codex_ask for asking questions, so it lacks explicit when-not-to-use guidance.

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

• codex_cloud_tasksA
  4.5/5.0
  Behavior5/5

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

  Annotations already declare readOnlyHint=true, destructiveHint=false, idempotentHint=true, openWorldHint=true, which covers the safety profile. The description adds valuable behavioral context: it wraps undocumented CLI commands, warns that field names in list output are not guaranteed stable, and notes that status output is raw (no --json form). These details go beyond annotations and help the agent set expectations correctly.

  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?

  The description is concise and front-loaded with the primary action ('List... or show status'). Every sentence earns its place: the first states purpose, the second explains the CLI wrapper and instability caveat, the third clarifies the status output format. No fluff or 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?

  With no output schema, the description does a good job of explaining return behavior: list returns JSON (with caveat about field stability), status returns raw output. It covers the tool's scope and limitations (undocumented CLI), and the schema handles parameter details. The tool is simple, and the description is complete enough for an agent to invoke it correctly.

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

  Parameters3/5

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

  Schema description coverage is 100%, with each parameter already well-documented (e.g., 'Ignored when task_id is set'). The description adds some context about task_id switching to status mode and the raw status output, but the schema already conveys this. Given the baseline of 3 for high schema coverage, this is appropriate; the description does not materially improve parameter understanding.

  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 clearly states the tool's verb ('List' / 'show the status of one') and resource ('Codex Cloud background tasks'). It distinguishes itself from siblings by focusing specifically on Codex Cloud tasks, and even mentions the wrapping of undocumented CLI commands. This is a specific, well-defined purpose.

  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?

  The description gives clear context for when to use the tool (listing tasks or showing status via task_id) and implicitly differentiates from siblings (codex_ask, codex_local_sessions, codex_read_thread). It doesn't explicitly state when not to use it, but the scope is evident. A minor gap is not naming alternative tools for other task-related operations, but the context is sufficient.

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

• codex_askA
  4.6/5.0
  Behavior5/5

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

  The description exceeds the annotations by disclosing that Codex can run shell commands and edit files when sandbox_mode is workspace-write, and that calls with different/omitted thread_id run concurrently. This adds operational context beyond the simple readOnly/destructive hints and aligns with the annotations, showing no contradiction.

  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?

  The description is compact—three sentences—with the most important information front-loaded: the core action, followed by key behavioral caveats. No redundant or filler content.

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

  Completeness4/5

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

  Given the tool's complexity (13 parameters) and rich schema, the description covers the essential behavioral aspects: prompt sending, thread continuity, concurrency, and file-edit capability. It doesn't describe the response format, but without an output schema and with detailed parameter docs, this is acceptable.

  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 schema already documents all 13 parameters with 100% coverage, so the baseline is 3. The description adds value by explaining concurrency behavior and the practical effect of sandbox_mode, which complements the schema though most param details are already covered.

  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 specific verb+resource ('Send a prompt to OpenAI Codex') and clearly explains the tool's scope, including its ability to execute commands and edit files under certain sandbox modes. It distinguishes itself from siblings by focusing on prompting Codex, while siblings like codex_read_thread and codex_local_sessions clearly handle other concerns.

  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?

  The description provides clear usage patterns: pass thread_id to continue a conversation, omit it to start fresh, and use independent calls for parallel queries. While it doesn't explicitly contrast with sibling tools, it gives sufficient context on when to use this tool and how to manage conversations.

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

• codex_local_sessionsA
  4.7/5.0
  Behavior5/5

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

  Adds significant context beyond annotations: reads an undocumented index file directly, exposes only three specific fields, and does not provide live/idle status or working directory. Annotations already say readOnlyHint/idempotentHint/destructiveHint, and the description expands on the data source and limitations without contradiction.

  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 compact sentences: action, rationale, limitations. No redundant fluff, every sentence adds value.

  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 simple read-only list tool with one optional parameter and no output schema, the description sufficiently explains purpose, data source, return contents, and limitations, making it complete for the agent to select and invoke correctly.

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

  Parameters3/5

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

  Schema coverage is 100% with `limit` fully described (default, min, max, ordering). The description adds no new parameter details, but the schema handles it; baseline 3 is appropriate.

  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?

  Clearly states it lists recent local Codex CLI sessions, specifies the source (sessions `codex resume` picks from), and distinguishes from siblings by noting it reads the session_index.jsonl index directly and what fields it exposes.

  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?

  Explains when to use this tool: to get a scriptable list of sessions for `codex resume`, and why not to use `codex resume` itself (interactive-only). It also sets exclusions: doesn't expose live/idle status or working directory, guiding the agent to not expect those.

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