omnigent-ai/omnigent

API Docs generates reference documentation for public APIs by reading the actual source code signatures, parameters, defaults, return shapes, and raised errors. Use it when you need accurate API reference or usage docs for a module, class, CLI, or endpoint.

Antigravity Native E2E Testing spins up a live local Omnigent server and runner, launches the real agy TUI, and exercises the antigravity-native harness end-to-end through web UI turns. Use it when developing, testing, or debugging the antigravity-native harness and its agy launch, RPC mirror, tmux delivery, OAuth, or MCP-relay behavior.

Antigravity SDK E2E Dev spins up a live local Omnigent server and exercises the Antigravity (Gemini) SDK harness end-to-end, including building agents, running turns, and smoke-testing. Use it when developing, testing, or debugging the Antigravity harness.

Build Omnigent Agents offers patterns and templates for creating valid Omnigent agent directories, covering minimal, web-search, harness, MCP, and multi-agent configurations. Use it when generating files for an Omnigent agent.

CLI Setup Verification drives the real Omnigent CLI through a pseudo-terminal in an isolated sandbox to verify setup/onboarding, terminal UI, and critical user journeys. Use it when developing or reviewing CLI setup/onboarding/REPL/picker changes, reproducing cold-start UX bugs, or confirming a fix lands via a before/after baseline diff.

Changelog turns a range of commits or merged PRs into a changelog entry grouped by change type. Use it when asked for release notes, a changelog, or what changed between two points.

Codex E2E Greet is a test greeter fixture for Codex skills end-to-end tests. It includes the unique suffix a3f9c2 in its skill name so a string match in the agent's output proves Codex loaded this skill.

Codex E2E XYZ Count B8d4e7 is a fixture skill used in end-to-end tests to verify that Codex enumerates skills. It includes a unique suffix so a string match confirms the skill was loaded.

Copilot SDK E2E Dev spins up a live local Omnigent server and exercises the Copilot SDK harness end-to-end, including building agents, running turns, smoke testing, and debugging auth, model, and tool-bridge behavior. Use it when developing or debugging the copilot executor, harness, or auth modules.

Cross-Review verifies an implementer's diff using an independent sub-agent from a different vendor, reviewing only the diff and contract. It turns blocking issues into fix-tasks and loops until gates are green and no blockers remain.

Cursor SDK Harness E2E spins up a live local Omnigent server and exercises the Cursor SDK harness end-to-end, building cursor agents and running real turns to smoke-test and bug-bash. Use it when developing, testing, or debugging the cursor harness or its auth, model, and tool-bridge behavior.

Debate has the Claude and GPT partners critique each other's answers across a configurable number of rounds (default 1) before converging on a synthesis. Use when you want the two perspectives stress-tested against each other, not just shown side by side.

Deep Research answers a question with a thorough, cross-checked, cited web-research report using search and fetch tools. Use it when you need current, verifiable information from the web.

Deploy Docker Compose brings up the Omnigent server as a Docker compose stack with Postgres on any Docker host. Use it to build the image, run the stack, debug on an existing host, or extend the stack to a new platform.

Fanout runs independent subtasks in parallel by creating a separate git worktree and dispatching an implementation sub-agent for each task, each opening its own PR. Use it when subtasks are parallel-safe and you want cross-reviewed PRs delivered without merging.

Framework Detection scans Python imports in existing agent code and maps them to Omnigent executor types. Use it when integrating existing Python agent code into Omnigent.

Harness Integration Guide provides a reference for building new Omnigent harness integrations, covering SDK/subprocess and native harness tracks with feature matrices, implementation patterns, and prioritized checklists.

Investigate delegates read-only investigation, debugging, audit, search, and code-understanding tasks to sub-agents, then synthesizes findings only from their structured reports. Use it when an orchestrator needs to answer repository-specific technical questions or analyze failures without modifying files.

Migration Guide turns a breaking change (an API rename, removed flag, changed default, or moved module) into concrete upgrade steps with before/after examples. Use when you need to migrate, upgrade, or adapt to a breaking change.

Omnigent Knowledge is a deep reference on Omnigent config format, executor types, skill/tool structure, and conventions. Use it when you need to look up how the platform works.

Pi Native E2E Dev spins up a live local Omnigent server and runner to exercise the native Pi TUI harness end-to-end. Use it when developing, testing, or debugging the pi-native harness, its bridge, extension, auth, or model behavior.

Polly E2E Testing runs deterministic mock-LLM and live CLI checks against the polly multi-agent coding orchestrator's critical user journeys. Use it when developing, debugging, or testing examples/polly, its sub-agents, skills, or guardrails.

Run Load Test drives Omnigent's load-test stack to benchmark concurrency with real host-bound sessions and a mocked LLM. Use it when load-testing, stress-testing, or benchmarking Omnigent under concurrency.

Security Audit examines a codebase or directory for security issues such as hardcoded secrets, injection, unsafe deserialization, weak cryptography, and authorization gaps, then produces a structured findings report. Use it when asked for a security review or vulnerability audit; it reports findings only and never fixes code.