Reference

The CLI, the MCP tools, configuration and the API, much of it generated from the code.

• Which CLIs and keys do I need?

  What each nexus-agents feature needs (a coding CLI, an API key or a gateway), which CLIs serve voter seats, and what the doctor statuses mean.

  • Reference
  • For users

• CLI reference

  Generated reference for every nexus-agents CLI command, audience, and description.

  • Reference
  • For users

• Configuration guide

  Configure nexus-agents with YAML files, environment variables, and programmatic options

  • Reference
  • For users

• Environment variable reference

  Registered NEXUS environment variables, accepted values, schema defaults and descriptions.

  • Reference
  • For users

• Nexus-agents entrypoints

  Complete reference for all CLI commands, MCP tools, and workflow templates

  • Reference
  • For users

• MCP tool reference

  Per-tool reference for all 47 registered nexus-agents MCP tools, generated from the tool manifest and input schemas.

  • Reference
  • For users

• MCP tool: cancel_job

  Cancel an async-mode job; aborts in-flight voters and workers — idempotent (#3042)

  • Reference
  • For users

• MCP tool: ci_health_check

  CI infrastructure health — composes GitHub status + recent-runs activity (#3076)

  • Reference
  • For users

• MCP tool: compare_data_feeds

  Diff two YAML/JSON feeds: coverage + per-field axes

  • Reference
  • For users

• MCP tool: consensus_vote

  Multi-model consensus voting on proposals

  • Reference
  • For users

• MCP tool: create_expert

  Create a specialized expert agent

  • Reference
  • For users

• MCP tool: delegate_to_model

  Pick the best-fit existing model for a task (no registry change)

  • Reference
  • For users

• MCP tool: execute_expert

  Run a task through a previously-created expert (by expertId)

  • Reference
  • For users

• MCP tool: execute_spec

  Execute AI software factory spec pipeline

  • Reference
  • For users

• MCP tool: extract_symbols

  TypeScript-compiler-API AST symbols from a SINGLE file (functions/classes/types)

  • Reference
  • For users

• MCP tool: get_job_result

  Read result of an async-mode dispatch by jobId (#3042 / #2631)

  • Reference
  • For users

• MCP tool: improvement_review

  Threshold-gated observability loop — surfaces routing/tech-debt/bug/security signals from outcome+fitness data; files candidate issues

  • Reference
  • For users

• MCP tool: issue_triage

  Triage GitHub issues with trust classification

  • Reference
  • For users

• MCP tool: list_available_models

  Probe all model-discovery transports (OpenRouter API + opencode/claude/codex/gemini CLIs) and report per-transport health — validates the CLIs/APIs are reachable (#3406)

  • Reference
  • For users

• MCP tool: list_experts

  Inventory of expert ROLES for `create_expert`

  • Reference
  • For users

• MCP tool: list_jobs

  List async-mode jobs across all tools — cross-session discovery (#3046 / #2631)

  • Reference
  • For users

• MCP tool: list_workflows

  Inventory of multi-step TEMPLATES for `run_workflow`

  • Reference
  • For users

• MCP tool: memory_query

  Query across all memory backends

  • Reference
  • For users

• MCP tool: memory_stats

  Memory system statistics dashboard

  • Reference
  • For users

• MCP tool: memory_write

  Write to typed memory backends

  • Reference
  • For users

• MCP tool: orchestrate

  Task orchestration with Orchestrator coordination

  • Reference
  • For users

• MCP tool: pr_review

  Multi-voter PR review with verification gate (experimental)

  • Reference
  • For users

• MCP tool: query_task_state

  Query the structured task-state log for a task ID

  • Reference
  • For users

• MCP tool: query_trace

  Query execution traces for observability

  • Reference
  • For users

• MCP tool: registry_import

  Draft YAML for a NEW model entry (for picking existing models use `delegate_to_model`)

  • Reference
  • For users

• MCP tool: repo_analyze

  Analyze GitHub repository structure

  • Reference
  • For users

• MCP tool: repo_security_plan

  Generate security scanning pipeline for a repo

  • Reference
  • For users

• MCP tool: research_add

  Add an arXiv PAPER to the registry (for non-paper sources use `research_add_source`)

  • Reference
  • For users

• MCP tool: research_add_source

  Add a NON-PAPER source (repo/tool/blog) — for arXiv papers use `research_add`

  • Reference
  • For users

• MCP tool: research_analyze

  Analyze registry for gaps, trends, coverage

  • Reference
  • For users

• MCP tool: research_catalog_review

  Review auto-cataloged research references

  • Reference
  • For users

• MCP tool: research_discover

  Discover papers/repos from external sources

  • Reference
  • For users

• MCP tool: research_query

  Query research registry (status, overlap, stats, search)

  • Reference
  • For users

• MCP tool: research_synthesize

  Synthesize registry into topic clusters with themes

  • Reference
  • For users

• MCP tool: run

  Default entry point — give a goal, MetaOrchestrator picks the strategy; returns the routing decision (execute:false, read-only) or runs it inline (execute:true; dev-pipeline+pipeline+research+consensus wired) (#3548)

  • Reference
  • For users

• MCP tool: run_dev_pipeline

  Full dev pipeline: research, plan, vote, implement, QA

  • Reference
  • For users

• MCP tool: run_graph_workflow

  Run a DAG workflow with per-node checkpoints + audit trail (linear → `run_workflow`)

  • Reference
  • For users

• MCP tool: run_pipeline

  Execute a pipeline plugin by name with typed input

  • Reference
  • For users

• MCP tool: run_quality_gate

  Run the QA quality gate (typecheck/lint/tests/build/security) over a project dir; returns structured pass/fail verdict + feedback

  • Reference
  • For users

• MCP tool: run_workflow

  Run a linear workflow template (use `run_graph_workflow` for DAGs)

  • Reference
  • For users

• MCP tool: search_codebase

  Cross-file search over declared symbol NAMES (declarations only, not usages)

  • Reference
  • For users

• MCP tool: search_usages

  Structural usage/call-site search for a symbol via ast-grep (calls, member calls, new, imports, references) — the "where is X used" gap `search_codebase` cannot fill

  • Reference
  • For users

• MCP tool: suggest_research_tasks

  SUGGEST-ONLY: candidate pipeline tasks from research_discover findings for review — files/executes nothing (#1715)

  • Reference
  • For users

• MCP tool: supply_chain_tradeoff_panel

  Per-axis tradeoff vote for build-vs-buy / supply-chain decisions

  • Reference
  • For users

• MCP tool: survey_oss_landscape

  Transient OSS project search (license, stars, last-commit) via GitHub

  • Reference
  • For users

• MCP tool: vendor_publishing_audit

  Look up a vendor's signing infrastructure (GPG keys, URL patterns, signature shape)

  • Reference
  • For users

• MCP tool: verify_audit_chain

  Verify hash chain of a FileAuditStorage audit log directory

  • Reference
  • For users

• MCP tool: weather_report

  Multi-CLI performance weather report

  • Reference
  • For users

• Strategy reference (force-strategy escape hatches)

  Manifest-generated reference for all 8 routable execution strategies: entrypoint tool, when to force, maturity/authority tier, executor availability.

  • Reference
  • For users

• API reference

  Generated from the TypeScript source with TypeDoc: every public export.

• API key usage boundaries

  Which API keys can be used with which CLI tools

  • Reference
  • For users

• Price basis in recorded costs

  Meaning and persisted-reader compatibility of the published PriceBasis vocabulary

  • Reference
  • For users

• Rule-loading precedence across harnesses

  Exactly which rule files Claude Code / Codex / Gemini CLI / OpenCode load, from where, in what order

  • Reference
  • For users