earakely-scale/NoaaTidesEnv-REA

noaa-tides

noaa-tides

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

Tables

• noaa-tides_tasks: one row per task of the bundle, with its prompt, envs, verifiers and steps.
• noaa-tides_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/noaa-tides.jsonl holds each run's record and native trajectory, and bundles/noaa-tides/ the bundle itself.

Run it

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

NOAA Tides and Currents API on AgentEnv, from a REA capture

What this is. A replay environment of NOAA Tides and Currents API v1 (U.S. public domain), built from one recorded browser session against the public API at https://api.tidesandcurrents.noaa.gov on 2026-10-19. The responses are real public-domain records published by NOAA's Center for Operational Oceanographic Products and Services, frozen as they were served then. REA projected the capture into an environment manifest (12 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 NOAA Tides and Currents API. Tasks are graded deterministically, with no judge.

What's in it

• 12 tools (12 GET) over 56 recorded exchanges on https://api.tidesandcurrents.noaa.gov; no recorded group excluded.
• Path parameters: 8 numeric, 1 sibling-variance. 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 21 tools.
• Unknowns the manifest states explicitly: authorization model, response semantics beyond recorded exchanges, state transitions between tools.
• 9 retrieval tasks (evals/noaa-tides-retrieval.toml): each asks one factual question about a recorded record, with a distinct answer, 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.
• 18 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_5ca46908fb465b0c7dddd2fa0a37a3f72870fa0add96c1ae818a5b9cb36c4766 (scoped HAR sha256 7e8a15258907151dac9391ac8687a3b17916c9a78208841634f11a33934e948f, 311,519 bytes), manifest view ev_a54a45f249d0ea4883e3b687ab1e4afd9c1671b34e14226298aeb2542dbdad82, confidence inferred, authority historical-reference. The REA build is a prerelease of the environment commands proposed upstream in morluto/rea#4954, #4955, #4975 and #4999; 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 behave like the API did on 2026-10-19. The live API keeps changing, so every answer is about the recorded snapshot, not today, and a query the session never made falls back to the closest record. The API is read-only and so is the environment: it has no state and no write tools. The manifest's unknowns say what replay cannot know.

Tasks and grading

task	tool	arguments	field	answer
qa-01-get_mdapi_prod_webapi_stations_sensors_b	get_mdapi_prod_webapi_stations_sensors_by_id	id=9414290	sensors[0].refdatum	Site Elevation
qa-02-get_mdapi_prod_webapi_stations_by_statio	get_mdapi_prod_webapi_stations_by_station_id	station_id=9414290.json	stations[0].name	San Francisco
qa-03-get_mdapi_prod_webapi_stations_benchmark	get_mdapi_prod_webapi_stations_benchmarks_by_id	id=8410140	benchmarks[0].designation	TBM 841 0140 RM OF ETG
qa-04-get_mdapi_prod_webapi_stations_harcon_by	get_mdapi_prod_webapi_stations_harcon_by_id	id=9414290, units=metric	HarmonicConstituents[0].description	Principal lunar semidiurnal constituent
qa-05-get_mdapi_prod_webapi_stations_datums_by	get_mdapi_prod_webapi_stations_datums_by_id	id=9414290, units=english	datums[0].description	Station Datum
qa-06-get_mdapi_prod_webapi_stations_products_	get_mdapi_prod_webapi_stations_products_by_id	id=9447130	products[0].name	Water Levels
qa-07-get_dpapi_prod_webapi_htf_htf_annual	get_dpapi_prod_webapi_htf_htf_annual	station=8638610, year=2024	AnnualFloodCount[0].stnName	Sewells Point, VA
qa-08-get_mdapi_prod_webapi_stations	get_mdapi_prod_webapi_stations	type=airgap	stations[0].name	Verrazano-Narrows Air Gap
qa-09-get_mdapi_prod_webapi_stations_datums_by	get_mdapi_prod_webapi_stations_datums_by_id	id=9414290, units=english	accepted	Apr 17 2003

Episodes

model	tasks	passed	pass rate	mean tool calls	mean input tokens
openai/gpt-oss-120b:groq	9	9	100%	1.0	5,895
zai-org/GLM-5.3-Flash:zai-org	9	9	100%	1.0	7,138

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/NoaaTidesEnv-REA --bundle noaa-tides --task smoke          # Docker, no model
agent-env hf run earakely-scale/NoaaTidesEnv-REA --bundle noaa-tides --eval noaa-tides-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. No instance was run: a throwaway VM called the public API directly, within the publisher's fair-access rules (a declared User-Agent, the application parameter on data requests, one request at a time and at most 2 requests per second (about 72 requests in the session)).
2. Playwright opened the API origin and ran a scripted sequence of read requests (station records for ten named water-level stations (San Francisco, Key West, Seattle, Galveston Pier 21, Sewells Point, San Diego, Eastport, Anchorage, Grand Isle, Hilo) with their details, datums, harmonic constituents, sensors, flood levels, notices, products and benchmarks, the air-gap station list, short-range tide predictions, water levels, monthly means, high/low and weather readings, and two annual high-tide flooding counts); the HAR was scoped to the API, with 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	NOAA Tides and Currents API (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)	12 tools mirroring the application's endpoints
State	Full: a virtual clock, triggers, frozen windows, agents that answer	None: every call is independent; the API is read-only
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	Real public-domain records from NOAA's Center for Operational Oceanographic Products and Services, as served on 2026-10-19
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 of the public API at https://api.tidesandcurrents.noaa.gov, published by NOAA's Center for Operational Oceanographic Products and Services (U.S. public domain); the dataset repeats them as recorded. REA is MIT software by morluto; AgentEnv Framework is Apache-2.0 by Scale AI.