Generic Harness
The Generic harness is the recommended default for most use cases. It configures 16 capabilities: 14 user-facing defaults plus cross-cutting helpers for tool narration and side questions. Together they cover file operations, command execution, web access, memory, budgeting, context management, and durable tool output.
When to Use
Section titled “When to Use”- General-purpose assistants
- Coding and scripting tasks
- Research workflows
- Any session where you want a solid set of defaults
Configuration
Section titled “Configuration”| Property | Value |
|---|---|
| Type | generic |
| System Prompt | ”You are a helpful assistant.” |
| Default Model | None (inherits from agent or organization) |
Bundled Capabilities
Section titled “Bundled Capabilities”| Capability | What it provides |
|---|---|
| File System | Read, write, list, grep, and delete files in the session workspace (/workspace) |
| Virtual Bash | Sandboxed bash shell for running commands, scripts, and text processing |
| Web Fetch | Fetch web content with file download support |
| Storage | Key/value store for general data and encrypted secret storage |
| Session | Access session metadata and manage session title |
| Session Schedules | Create and manage cron-style schedules that wake the session |
| AGENTS.md | Reads AGENTS.md from workspace and injects project-level instructions |
| Agent Skills | Discover and activate skills from /.agents/skills/ |
| Infinity Context | Trims older messages from the live prompt while exposing earlier history via query_history |
| OpenAI Tool Search | Defers tool schema loading on supported models to reduce prompt size |
| Context Compaction | Auto-compacts context at 85% budget via cascading strategies |
| Budgeting | Token budget enforcement with configurable meters and rules |
| Self-Budget | Prompt-only guidance for reasoning about a user-requested indicative budget using session usage data |
| Tool Output Persistence | Persists full tool output to /.outputs/ before truncation for lossless retrieval |
Infinity Context and Context Compaction work together to keep long sessions unbounded. See Context Compaction for details.
See Also
Section titled “See Also”- Base Harness — empty harness for full control
- Platform Chat Harness — extends Generic with platform management
- Harnesses feature guide — harness selection and API management