earakely-scale/AirbyteEnv-REA

airbyte

airbyte

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

Tables

• airbyte_tasks: one row per task of the bundle, with its prompt, envs, verifiers and steps.
• airbyte_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/airbyte.jsonl holds each run's record and native trajectory, and bundles/airbyte/ the bundle itself.

Run it

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

Airbyte on AgentEnv, from a REA capture

What this is. An environment of Airbyte 2.4.0 (Elastic License 2.0 (ELv2)), built from one recorded browser session against a self-hosted instance seeded with synthetic data. REA projected the capture into an environment manifest (25 tools) and an environment state, the entities the capture's reads showed, and rea environment-export wrote the bundle you can run from this dataset. Reads the state can render are answered from it, write tools whose effect the capture confirmed change it, and data/reset restores it. Nothing executes Airbyte. Tasks are graded deterministically, with no judge.

v0.2.0 rebuilds this environment from the same capture with a REA build that serves a state, confirmed writes and list queries; the tag v0.1.0 keeps the earlier replay-only version.

What's in it

• 25 tools (18 GET, 3 POST, 2 DELETE, 2 PATCH) over 64 recorded exchanges on http://127.0.0.1:3302; no recorded group excluded. 23 are offered; the other 2 are write tools the environment cannot apply (post_public_jobs, delete_public_sources_by_uuid), left out rather than faked.
• Path parameters: 14 uuid, 1 numeric. REA's four format rules (numeric, uuid, hex, iso-date) applied unconditionally; the four opt-in rules (after-uid-literal, opaque-token, sibling-variance, prefix-variance) were requested and are labelled per parameter. With the four format rules alone the manifest has 25 tools.
• State: 73 entities of 9 types. 17 read tools answer from it and 1 replay their recorded responses. 6 list tools also answer filters, search, sorting or paging from it. Served write tools: post_public_sources (create), patch_public_sources_by_uuid (update), post_public_tags (create), delete_public_tags_by_uuid (delete), patch_public_tags_by_uuid (update).
• Unknowns the manifest states explicitly: authorization model, response semantics beyond recorded exchanges, state transitions between tools.
• 11 retrieval tasks (evals/airbyte-retrieval.toml): each asks one factual question about a seeded object, with a distinct answer that the app does not ship by default; reward 1 if the response contains the recorded value, else 0. Plus smoke, which deploys the environment and checks it through the MCP endpoint.
• 22 retrieval episodes, plus the smoke run, 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_8e00dfaee9452108d545ea793346d937d5407358ab2ea30d6d614fcfac225307 (scoped HAR sha256 1d67ea816437d694c369791740572fbab7d10aa6826ee568c5f98c7608e32016, 124,694 bytes), manifest view ev_bee052026ce2bf4e8ad3a304714615f99105ae607a14fab7cf0b45ab486dd498, confidence inferred, authority historical-reference. The REA build is rea-agents 6.7.0 from main 335c042d with open pull requests merged: morluto/rea#5600 at fe8ab584 (which carries #5351 and #5352), #5664 at 03b66df9, #5787 at edc09dba, #5800 at 6db41d29 and #5699 at ac3ca8a3. The bundle installs it from the tarball it ships, so the environment runs without waiting for an npm release.

What the state means, and does not

Reads the state can render, item, list and document views of the entities the capture's responses showed, answer from it. List reads filtered, searched, sorted or paged answer from it too wherever the capture's recorded calls explain what each query parameter does, including values no call recorded; a parameter the recordings do not explain, or a flag's unrecorded value, replays. Other reads replay the recorded response whose path and query values best match the arguments, with replay.exact_match saying whether it matched exactly; a replayed read does not change after a write. A served write was confirmed by the capture itself: replaying the session against the extracted state, a later recorded read reproduced its effect and none contradicted it. A create generates ids and clock values deterministically, so the same calls after a reset give the same state; an update changes the members the call sends; a delete removes the entity. Each takes only the body members the recorded call showed and refuses others with a message that names the accepted ones. Its effect shows in the reads the state renders, lists included wherever a list lists its whole collection or a learned query answers the call; a list call nothing explains keeps its recorded members, dropping a deleted entity but not adding a new one. The data plane at /agentenv takes data/reset (restore the extracted state), data/add and data/get (return it), which is how write tasks are graded.

Not modelled: side effects a write has beyond its own entity (counters, copies in other collections, related entities), preconditions the application enforces before a write, list query parameters the recorded calls do not explain, paging members such as page counts and next links when a call asks for another page size or page, and anything the capture never showed. The manifest's unknowns list what a capture cannot establish.

Tasks and grading

Retrieval

task	tool	arguments	field	answer
qa-01-get_public_connections_by_uuid	get_public_connections_by_uuid	uuid=68ba2580-878c-478d-b080-8f2712ac28c4	name	Saltmarsh Depot tide readings relay
qa-02-get_public_workspaces_by_uuid	get_public_workspaces_by_uuid	uuid=4cfeabc5-3805-40d8-ae6c-436f584f2367	name	Saltmarsh Depot Sandbox
qa-03-get_public_destinations_by_uuid	get_public_destinations_by_uuid	uuid=bf35d114-580d-43be-bf03-8cdea60ec161	name	Lighthouse Rota Quiet Drain
qa-04-get_public_tags_by_uuid	get_public_tags_by_uuid	uuid=dfba26d0-7573-464a-a5ac-471fa19a898f	name	Kitchen Night Shift
qa-05-get_public_sources_by_uuid	get_public_sources_by_uuid	uuid=78805f75-f349-4cc3-a43d-f54709cbcca3	name	Copperfield Annex Visitor Log
qa-06-get_public_workspaces	get_public_workspaces	no arguments	data[0].name	Lighthouse Rota Pipelines
qa-07-get_public_organizations	get_public_organizations	no arguments	data[0].organizationName	Lighthouse Rota
qa-08-get_public_tags	get_public_tags	no arguments	data[0].name	Quarry Audit
qa-09-get_public_connections	get_public_connections	no arguments	data[0].name	Blue Heron Kitchen nightly orders feed
qa-10-get_public_destinations	get_public_destinations	no arguments	data[0].name	Copperfield Annex Null Sink
qa-11-get_public_sources	get_public_sources	no arguments	data[0].name	Blue Heron Kitchen Orders

Episodes

model	retrieval episodes	passed	pass rate	mean tool calls	mean input tokens
openai/gpt-oss-120b:groq	11	9	82%	1.0	12,699
zai-org/GLM-5.3-Flash:zai-org	11	11	100%	1.0	12,795

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

gpt-oss-120b:groq missed qa-09-get_public_connections, qa-10-get_public_destinations.

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/AirbyteEnv-REA --bundle airbyte --task smoke          # Docker, no model
agent-env hf run earakely-scale/AirbyteEnv-REA --bundle airbyte --eval airbyte-retrieval --model openai/gpt-oss-120b:groq

The evals need 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. Airbyte ran in Docker on a throwaway VM, bound to loopback, seeded through its own API with synthetic records (two workspaces (Lighthouse Rota Pipelines, Saltmarsh Depot Sandbox), four Sample Data sources (Blue Heron Kitchen Orders, Granite Ledger Stone Census, Copperfield Annex Visitor Log, Saltmarsh Depot Tide Readings), three End-to-End Testing (/dev/null) destinations (Copperfield Annex Null Sink, Lighthouse Rota Quiet Drain, Saltmarsh Depot Dev Null), three connections, three tags (Quarry Audit, Kitchen Night Shift, Tide Table Crew) and one finished sync job (900 rows)); 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 and the state it serves; 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	Airbyte (this dataset)
Environment	A simulation: a container quay, ships, cranes, weather, with A2A ship agents	An application's API rebuilt from one captured session
Observation	Tool results computed by the simulator as the week unfolds	Answers rendered from the extracted state, or recorded responses, cited to Evidence
Action space	Planning tools (check_plan, confirm_berths, advance, message_agent)	23 tools mirroring the application's endpoints
State	Full: a virtual clock, triggers, frozen windows, agents that answer	A state extracted from the capture (73 entities); served writes change it; data/reset restores it
Reward	Deterministic, against a CP-SAT optimum or a re-planner anchor; continuous 0 to 1	Deterministic: recorded-value match for retrieval; 0 or 1 per task
What it measures	Planning under uncertainty and negotiation	Tool discovery, parameter inference and faithful retrieval in a real API, and changing its state
Data provenance	Real 2024 port calls and weather; simulated workloads	Synthetic seed data on a self-hosted Airbyte; 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 and the state's limits above say what it still leaves out.

Licence and credits

Bundle code and tasks: Apache-2.0. Recorded content: responses generated by a self-hosted Airbyte instance over synthetic data; Airbyte is Elastic License 2.0 (ELv2) software by Airbyte, Inc. 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.