| Accountable human handoff | 0 / 31 | 31 | 0% |
| Administrative policy controls | 0 / 31 | 31 | 0% |
| Agent Plugins failure isolation | 0 / 31 | 31 | 0% |
| Agent Skills manual invocation | 0 / 31 | 31 | 0% |
| Agent Skills progressive loading | 0 / 31 | 31 | 0% |
| Agent Skills references and assets | 0 / 31 | 31 | 0% |
| Agent Skills scripts | 0 / 31 | 31 | 0% |
| AGENTS.md root instructions | 0 / 31 | 31 | 0% |
| Artifact and workspace export | 0 / 31 | 31 | 0% |
| Audit logs and actor attribution | 0 / 31 | 31 | 0% |
| Automatic context compaction | 0 / 31 | 31 | 0% |
| Concurrent subagents | 0 / 31 | 31 | 0% |
| Context compaction controls | 0 / 31 | 31 | 0% |
| Context usage visibility | 0 / 31 | 31 | 0% |
| Context window and effective input limit | 0 / 31 | 31 | 0% |
| Controlled conversation sharing | 0 / 31 | 31 | 0% |
| Conversation and run export | 0 / 31 | 31 | 0% |
| Data residency and processing location controls | 0 / 31 | 31 | 0% |
| Data retention and deletion controls | 0 / 31 | 31 | 0% |
| Encryption and key controls | 0 / 31 | 31 | 0% |
| Failure recovery and safe retry | 0 / 31 | 31 | 0% |
| File upload limits and quota visibility | 0 / 31 | 31 | 0% |
| Local models | 0 / 31 | 31 | 0% |
| Local-only and offline operation | 0 / 31 | 31 | 0% |
| MCP cancellation | 0 / 31 | 31 | 0% |
| MCP elicitation | 0 / 31 | 31 | 0% |
| MCP list-changed notifications | 0 / 31 | 31 | 0% |
| MCP logging | 0 / 31 | 31 | 0% |
| MCP OAuth authorization | 0 / 31 | 31 | 0% |
| MCP progress notifications | 0 / 31 | 31 | 0% |
| MCP prompt-list notifications | 0 / 31 | 31 | 0% |
| MCP prompts | 0 / 31 | 31 | 0% |
| MCP Registry metadata | 0 / 31 | 31 | 0% |
| MCP resource subscriptions | 0 / 31 | 31 | 0% |
| MCP resources | 0 / 31 | 31 | 0% |
| MCP revision compatibility | 0 / 31 | 31 | 0% |
| MCP roots | 0 / 31 | 31 | 0% |
| MCP sampling | 0 / 31 | 31 | 0% |
| MCP server instructions | 0 / 31 | 31 | 0% |
| MCP server role | 0 / 31 | 31 | 0% |
| MCP task listing | 0 / 31 | 31 | 0% |
| MCP tasks | 0 / 31 | 31 | 0% |
| Model fallback visibility and control | 0 / 31 | 31 | 0% |
| Model selection | 0 / 31 | 31 | 0% |
| Model training and improvement controls | 0 / 31 | 31 | 0% |
| Nested AGENTS.md instructions | 0 / 31 | 31 | 0% |
| Nested subagents | 0 / 31 | 31 | 0% |
| Output token limit | 0 / 31 | 31 | 0% |
| Per-subagent model selection | 0 / 31 | 31 | 0% |
| Plan and pricing disclosure | 0 / 31 | 31 | 0% |
| Plugin vendor extensions | 0 / 31 | 31 | 0% |
| Product-native rules | 0 / 31 | 31 | 0% |
| Project-scoped Agent Skills | 0 / 31 | 31 | 0% |
| Prompt and context cache reuse | 0 / 31 | 31 | 0% |
| Prompt cache controls | 0 / 31 | 31 | 0% |
| Prompt cache telemetry | 0 / 31 | 31 | 0% |
| Rate limit and quota disclosure | 0 / 31 | 31 | 0% |
| Reasoning effort controls | 0 / 31 | 31 | 0% |
| Resumable runs | 0 / 31 | 31 | 0% |
| Role-based collaboration access | 0 / 31 | 31 | 0% |
| Run timeout and duration limits | 0 / 31 | 31 | 0% |
| Run, approval, and failure notifications | 0 / 31 | 31 | 0% |
| Secrets management and redaction | 0 / 31 | 31 | 0% |
| Shared projects and team context | 0 / 31 | 31 | 0% |
| Structured subagent result handoff | 0 / 31 | 31 | 0% |
| Subagent approval and permission boundaries | 0 / 31 | 31 | 0% |
| Subagent context isolation and inheritance | 0 / 31 | 31 | 0% |
| Subagent lifecycle controls | 0 / 31 | 31 | 0% |
| Subagent MCP access | 0 / 31 | 31 | 0% |
| Subagent write access | 0 / 31 | 31 | 0% |
| Usage and cost metering | 0 / 31 | 31 | 0% |
| User-scoped Agent Skills | 0 / 31 | 31 | 0% |
| Video preprocessing disclosure | 0 / 31 | 31 | 0% |
| WebMCP execution cancellation | 0 / 31 | 31 | 0% |
| WebMCP permissions policy | 0 / 31 | 31 | 0% |
| WebMCP security annotations | 0 / 31 | 31 | 0% |
| Agent Plugins packaged MCP | 1 / 31 | 30 | 3% |
| Agent Plugins packaged skills | 1 / 31 | 30 | 3% |
| Agent Plugins portable manifest | 1 / 31 | 30 | 3% |
| Code execution | 1 / 31 | 30 | 3% |
| Custom instructions | 1 / 31 | 30 | 3% |
| Long-term memory | 1 / 31 | 30 | 3% |
| MCP client role | 1 / 31 | 30 | 3% |
| MCP legacy HTTP and SSE | 1 / 31 | 30 | 3% |
| MCP stdio transport | 1 / 31 | 30 | 3% |
| MCP Streamable HTTP | 1 / 31 | 30 | 3% |
| Native plugin system | 1 / 31 | 30 | 3% |
| Plugin marketplace and installation | 1 / 31 | 30 | 3% |
| Scheduled runs | 1 / 31 | 30 | 3% |
| Screenshots | 1 / 31 | 30 | 3% |
| Streaming output | 1 / 31 | 30 | 3% |
| WebMCP declarative form tools | 1 / 31 | 30 | 3% |
| WebMCP imperative tools | 1 / 31 | 30 | 3% |
| WebMCP tool discovery | 1 / 31 | 30 | 3% |
| WebMCP tool execution | 1 / 31 | 30 | 3% |
| Audio file upload and understanding | 2 / 31 | 29 | 6% |
| Background agents | 2 / 31 | 29 | 6% |
| Browser automation | 2 / 31 | 29 | 6% |
| Connectors | 2 / 31 | 29 | 6% |
| Human approval | 2 / 31 | 29 | 6% |
| Image upload and paste | 2 / 31 | 29 | 6% |
| PDF documents | 2 / 31 | 29 | 6% |
| Subagent delegation | 2 / 31 | 29 | 6% |
| Text and office document input | 2 / 31 | 29 | 6% |
| Video upload and understanding | 2 / 31 | 29 | 6% |
| MCP Apps | 4 / 31 | 27 | 13% |
| Computer use | 5 / 31 | 26 | 16% |
| Terminal commands | 7 / 31 | 24 | 23% |
| Workspace files | 8 / 31 | 23 | 26% |
| Agent Skills core and discovery | 9 / 31 | 22 | 29% |
| MCP tools | 10 / 31 | 21 | 32% |
