OpenProse Documentation

OpenProse is a portable, markdown-first workflow format for orchestrating AI sessions. In OpenClaw it ships as a plugin that installs an OpenProse skill pack and a /prose slash command. Programs live in .prose files and can spawn multiple sub-agents with explicit control flow.

Install, Turn on the OpenProse plugin and restart the Gateway.
Run a program, Run /prose run to execute a .prose file or remote program.
Write programs, Create multi-agent workflows with parallel and sequential steps.

Install

Enable the plugin

OpenProse is included but turned off by default. Enable it:

openclaw plugins enable open-prose

Restart the Gateway

openclaw gateway restart

Verify

openclaw plugins list | grep prose

You should see open-prose listed as enabled. The /prose skill command is now available in chat.

From a repo checkout you can install the plugin directly: openclaw plugins install ./extensions/open-prose

Slash command

OpenProse registers /prose as a user-invocable skill command:

/prose help
/prose run <file.prose>
/prose run <handle/slug>
/prose run <https://example.com/file.prose>
/prose compile <file.prose>
/prose examples
/prose update

/prose run <handle/slug> resolves to https://p.prose.md/<handle>/<slug>. Direct URLs are fetched as-is using the web_fetch tool.

Top-level remote runs are explicit. Remote imports inside a .prose program are transitive code dependencies: before OpenProse fetches any remote use target, it shows the resolved import list and requires the operator to reply exactly approve remote prose imports for that run.

What it can do

Multi-agent research and synthesis with explicit parallelism.
Repeatable, approval-safe workflows (code review, incident triage, content pipelines).
Reusable .prose programs you can run across supported agent runtimes.

Example: parallel research and synthesis

# Research + synthesis with two agents running in parallel.

input topic: "What should we research?"

agent researcher:
  model: sonnet
  prompt: "You research thoroughly and cite sources."

agent writer:
  model: opus
  prompt: "You write a concise summary."

parallel:
  findings = session: researcher
    prompt: "Research {topic}."
  draft = session: writer
    prompt: "Summarize {topic}."

session "Merge the findings + draft into a final answer."
  context: { findings, draft }

OpenClaw runtime mapping

OpenProse programs map to OpenClaw primitives:

OpenProse concept	OpenClaw tool
Spawn session / Task tool	`sessions_spawn`
File read / write	`read` / `write`
Web fetch	`web_fetch` (`exec` + curl when POST is needed)

Warning

If your tool allowlist blocks sessions_spawn, read, write, or web_fetch, OpenProse programs will fail. Check your tools allowlist config.

File locations

OpenProse keeps state under .prose/ in your workspace:

.prose/
├── .env                      # config (key=value), e.g. OPENPROSE_POSTGRES_URL
├── runs/
│   └── {YYYYMMDD}-{HHMMSS}-{random}/
│       ├── program.prose     # copy of the running program
│       ├── state.md          # execution state
│       ├── bindings/
│       ├── imports/          # nested remote program runs
│       └── agents/
└── agents/                   # project-scoped persistent agents

User-level persistent agents (shared across projects) live at:

~/.prose/agents/

State backends

filesystem (default)

State is written to .prose/runs/... in the workspace. No extra dependencies required.

in-context

Transient state kept in the context window; select with --in-context. Suitable for small, short-lived programs.

sqlite (experimental)

Select with --state=sqlite. Requires the sqlite3 binary on PATH (falls back to filesystem when missing); state lands in .prose/runs/{id}/state.db.

postgres (experimental)

Select with --state=postgres. Requires psql and a connection string in OPENPROSE_POSTGRES_URL (set it in .prose/.env).

Warning

Postgres credentials flow into sub-agent logs. Use a dedicated, least-privileged database.

Security

Treat .prose files like code. Review them before running, including remote use imports. Top-level /prose run https://... requests are explicit, but transitive remote imports require per-run approval before they are fetched or executed. Use OpenClaw tool allowlists and approval gates to control side effects. For deterministic, approval-gated workflows, compare with Lobster.

Skills reference, How OpenProse's skill pack loads and what gates apply.
Subagents, OpenClaw's native multi-agent coordination layer.
Text-to-speech, Add audio output to your workflows.
Slash commands, All available chat commands including /prose.

Official site: https://www.prose.md