earakely-scale/MailpitEnv-REA

mailpit

mailpit

An AgentEnv bundle with its recorded runs, published with agentenv-hf.

Tables

• mailpit_tasks: one row per task of the bundle, with its prompt, envs, verifiers and steps.
• mailpit_episodes: one row per recorded run, with its status, model, reward and scores, the prompt and response, and the agent's transcript as chat messages when its trajectory format is one agentenv-hf reads.

raw/mailpit.jsonl holds each run's record and native trajectory, and bundles/mailpit/ the bundle itself.

Run it

pip install agentenv-framework huggingface_hub
hf download earakely-scale/MailpitEnv-REA --repo-type dataset --local-dir agentenv-dataset
agent-env run agentenv-dataset/bundles/mailpit

Mailpit on AgentEnv, from a REA capture

What this is. A replay environment of Mailpit 1.31.6 (MIT), built from one recorded browser session against a self-hosted instance seeded with synthetic data. REA projected the capture into an environment manifest (15 tools), and rea environment-export wrote the bundle you can run from this dataset. Every tool call replays a retained record from the capture, cites the record and the Evidence it came from, and nothing executes Mailpit. Tasks are graded deterministically, with no judge.

What's in it

• 15 tools (9 GET, 3 PUT, 2 DELETE, 1 POST) over 68 recorded exchanges on http://127.0.0.1:3502; no recorded group excluded.
• Path parameters: 3 sibling-variance, 1 numeric. REA's four format rules (numeric, uuid, hex, iso-date) applied unconditionally; the three opt-in rules (after-uid-literal, opaque-token, sibling-variance) were requested and are labelled per parameter. With the four format rules alone the manifest has 46 tools.
• Unknowns the manifest states explicitly: authorization model, response semantics beyond recorded exchanges, state transitions between tools.
• 11 retrieval tasks (evals/mailpit-retrieval.toml): each asks one factual question about a seeded object, with a distinct answer that the app does not ship by default, answerable through a recorded exchange; reward 1 if the response contains the recorded value, else 0. Plus smoke, which deploys the environment and checks replay through the MCP endpoint.
• 22 recorded episodes by openai/gpt-oss-120b:groq, zai-org/GLM-5.3-Flash:zai-org through the Hugging Face router, with their chat transcripts.
• Provenance: capture ev_098d173ad7bdea45ef0a3553f70864e90172642c1184f7bfa2f583628448c6c6 (scoped HAR sha256 48daf9aaab8bb99a872ac520db76aaa0231e30ab1a6788b0fd65a125316b7091, 175,008 bytes), manifest view ev_7a9c155dab4ca3db1aa0526859d8138afe9d2bc49ce63659a0dbef3974df5690, confidence inferred, authority historical-reference. The REA build is rea-agents 6.7.0 from main at commit 3fca2e0c, which merged the environment commands (morluto/rea#4954, #4955, #4975, #4999 and #5182); the bundle installs it from the tarball it ships, so the environment runs without waiting for an npm release.

What replay means, and does not

A tool call returns the recorded response whose path and query values best match the arguments, with replay.exact_match telling you whether it matched exactly or fell back to the closest record. Reads of seeded objects behave like the application. Writes return the recorded answer of the one write the session made, and the environment has no state: a created object does not appear in a later list. This is what makes retrieval tasks hard-gradable and write tasks weak; the manifest's unknowns say so.

Tasks and grading

task	tool	arguments	field	answer
qa-01-get_message_by_message_id	get_message_by_message_id	message_id=6tlIZBnsU0d1p5ryN0GzDH	From.Name	Saltmarsh Depot dispatch
qa-02-get_message_part_by_message_id_id	get_message_part_by_message_id_id	message_id=6tlIZBnsU0d1p5ryN0GzDH, id=2	departure	dawn, Friday
qa-03-get_search	get_search	query=tag:copperfield-annex, limit=50	messages[0].From.Name	Odile Marsh
qa-04-get_message_latest	get_message_latest	no arguments	To[0].Name	Saltmarsh Depot weighbridge
qa-05-get_webui	get_webui	no arguments	Label	Saltmarsh Depot Outbox
qa-06-get_messages	get_messages	limit=50	messages[0].From.Name	Granite Ledger night clerk
qa-07-get_message_by_message_id	get_message_by_message_id	message_id=6tlIZBnsU0d1p5ryN0GzDH	Subject	Saltmarsh Depot pallet manifest SD-0417
qa-08-get_message_part_by_message_id_id	get_message_part_by_message_id_id	message_id=6tlIZBnsU0d1p5ryN0GzDH, id=2	cargo	Granite setts and kelp crates
qa-09-get_search	get_search	query=tag:copperfield-annex, limit=50	messages[0].Subject	Desk 4 at the Copperfield Annex needs a new lamp
qa-10-get_message_latest	get_message_latest	no arguments	Subject	Weighbridge tickets WB-7730 to WB-7736 for the night quarry run
qa-11-get_messages	get_messages	limit=50	messages[0].From.Address	night-clerk@example.test

Episodes

model	tasks	passed	pass rate	mean tool calls	mean input tokens
openai/gpt-oss-120b:groq	11	11	100%	1.0	6,734
zai-org/GLM-5.3-Flash:zai-org	11	11	100%	1.0	7,646

Every episode is in the episodes config with its transcript; rewards are the deterministic verifier scores.

Every recorded episode passed.

Run it yourself

uv tool install agentenv-framework --with 'agentenv-hf @ git+https://github.com/earakely-scale/agentenv-hf-plugin@v0.3.0'
agent-env hf run earakely-scale/MailpitEnv-REA --bundle mailpit --task smoke          # Docker, no model
agent-env hf run earakely-scale/MailpitEnv-REA --bundle mailpit --eval mailpit-retrieval --model openai/gpt-oss-120b:groq

The retrieval eval needs a model endpoint: set [model] base_url and api_key in .agentenv/config.toml (the Hugging Face router works with any hf_ token). The bundle's agents/replay-qa is a small OpenAI-compatible tool-calling agent; any A2A agent can replace it.

How it was made

1. Mailpit ran in Docker on a throwaway VM, bound to loopback, seeded through its own API with synthetic records (messages: 19 emails sent over SMTP (plain, HTML+text, multipart with attachments; Cc, Bcc and Reply-To; tags from X-Tags) between Ada Finch, Rui Castellanos, Mina Okafor, Tomas Vey, Wren Halloway, Odile Marsh and the desks of Lighthouse Rota, Blue Heron Kitchen, Saltmarsh Depot, Copperfield Annex and Granite Ledger, all @example.test; subjects examples: Granite Ledger quarterly stone count GL-2291, Saltmarsh Depot pallet manifest SD-0417, Desk 4 at the Copperfield Annex needs a new lamp, Weighbridge tickets WB-7730 to WB-7736 for the night quarry run; attachments: 7 small attachments: one...); people and projects are invented.
2. Playwright drove a login, the main pages and a scripted read, write, read sequence through the app's JSON API; the HAR was scoped to the API, the login exchange dropped, secret-looking body fields scrubbed, headers and cookies removed.
3. rea inspect-web-network-capture retained the exchanges as Evidence; rea inspect-analysis-view projected the environment-manifest view; rea environment-export wrote this bundle; gen_tasks.py derived the retrieval tasks from the manifest and the recorded bodies; agent-env run recorded the episodes; agent-env hf publish wrote this dataset. Scripts: https://github.com/earakely-scale/rea-environments.

Compared with PortSimEnv on AgentEnv

	PortSimEnv-AgentEnv	Mailpit (this dataset)
Environment	A simulation: a container quay, ships, cranes, weather, with A2A ship agents	A replay of a real application's API, captured from one session
Observation	Tool results computed by the simulator as the week unfolds	Recorded responses, cited to Evidence
Action space	Planning tools (check_plan, confirm_berths, advance, message_agent)	15 tools mirroring the application's endpoints
State	Full: a virtual clock, triggers, frozen windows, agents that answer	None: every call is independent; writes replay one recorded answer
Reward	Deterministic, against a CP-SAT optimum or a re-planner anchor; continuous 0 to 1	Deterministic, recorded-value match; 0 or 1 per task
What it measures	Planning under uncertainty and negotiation	Tool discovery, parameter inference and faithful retrieval in a real API
Data provenance	Real 2024 port calls and weather; simulated workloads	Synthetic seed data on a self-hosted Mailpit; nothing real
How it was built	Hand-written plugin, months of work	Generated from a capture in one scripted pass
Where it is strong	Long-horizon decisions with hard rewards	Breadth: any app you can capture becomes an environment

The two are complementary: PortSimEnv shows what a hand-built simulation can grade; this dataset shows how far a capture alone gets you, and the manifest's unknowns list exactly what a stateful version would still have to specify.

Licence and credits

Bundle code and tasks: Apache-2.0. Recorded content: responses generated by a self-hosted Mailpit instance over synthetic data; Mailpit is MIT software by Ralph Slooten and the Mailpit contributors and is not redistributed here beyond the recorded JSON responses. REA is MIT software by morluto; AgentEnv Framework is Apache-2.0 by Scale AI.