earakely-scale/EspoCRMEnv-REA

espocrm

espocrm

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

Tables

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

Run it

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

EspoCRM on AgentEnv, from a REA capture

What this is. An environment of EspoCRM 10.0.10 (AGPL-3.0), built from one recorded browser session against a self-hosted instance seeded with synthetic data. REA projected the capture into an environment manifest (88 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 EspoCRM. Tasks are graded deterministically, with no judge.

v0.2.0 rebuilds this environment from a new capture of an instance seeded the same way, recorded so each write's effect is shown: two scratch objects per write target, every write read back at once. A REA build that serves a state, confirmed writes and list queries projects it; the tag v0.1.0 keeps the earlier replay-only version.

What's in it

• 88 tools (85 GET, 1 POST, 1 DELETE, 1 PUT) over 147 recorded exchanges on http://127.0.0.1:3216; no recorded group excluded. 86 are offered; the other 2 are write tools the environment cannot apply (post_task, delete_task_by_uid), left out rather than faked.
• Path parameters: 35 opaque-token. 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 114 tools.
• State: 129 entities of 41 types. 85 read tools answer from it and 0 replay their recorded responses. 22 list tools also answer filters, search, sorting or paging from it. Served write tools: put_task_by_uid (update).
• Unknowns the manifest states explicitly: authorization model, response semantics beyond recorded exchanges, state transitions between tools.
• 12 retrieval tasks (evals/espocrm-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.
• 2 write tasks (evals/espocrm-writes.toml): each asks for a change in plain language and is graded from the environment's state through data/get, not from the agent's reply; reward 1 only if every check passes. Their grade is in scores.state; reward holds the retrieval verifier's score, so it is empty for them, as for smoke.
• 24 retrieval episodes and 4 write 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_fb0a231df418616cc524f959854fa255c18bec65cc5001f93f7f5e8763b8084c (scoped HAR sha256 f65da5bb7426065f8049ab2016bbfdf8759e28b7d18e4eeb5514c50c5bfa527f, 316,114 bytes), manifest view ev_86e657936671a41c320c86869c29e738e15e3bf77fa5b217a4f59a5bf2b5920f, confidence inferred, authority historical-reference. The REA build is rea-agents 6.7.0 from main 3e7e102a with open pull requests merged: morluto/rea#6385 at 342933ea and morluto/rea#6897 at 08c2532a. 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_campaign_by_uid	get_campaign_by_uid	uid=6adb96396db47d7a9	description	Postcards to every keeper on the Lighthouse Rota.
qa-02-get_team_by_uid	get_team_by_uid	uid=6adb962aa7a7ad5ee	name	Lighthouse Rota Crew
qa-03-get_meeting_by_uid	get_meeting_by_uid	uid=6adb9636682ba2882	name	Granite Ledger quarterly review
qa-04-get_user_by_uid	get_user_by_uid	uid=6adb961384d6729ec	title	CRM Administrator
qa-05-get_task_by_uid	get_task_by_uid	uid=6adb9634d11e277f8	description	Include the five-year plan and seal 77-B notes.
qa-06-get_contact_cases_by_uid	get_contact_cases_by_uid	uid=6adb962dafbdb7d15, maxSize=5, offset=0, orderBy=number, order=desc, attributeSelect=name,number,status,priority	list[0].name	Weighbridge 2 reads 40 kg heavy
qa-07-get_knowledgebasearticle_by_uid	get_knowledgebasearticle_by_uid	uid=6adb96386f79dad99	name	Calibrating a weighbridge after heavy rain
qa-08-get_opportunity_by_uid	get_opportunity_by_uid	uid=6adb9631452a8984e	name	Granite Ledger quarry scales renewal
qa-09-get_call_by_uid	get_call_by_uid	uid=6adb96376fa5889b1	name	Call Tomas Vey about drop schedule
qa-10-get_targetlist_by_uid	get_targetlist_by_uid	uid=6adb9639e63c562d8	description	Keepers and cooks who receive the winter mailing.
qa-11-get_account_contacts_by_uid	get_account_contacts_by_uid	uid=6adb962ba77cd76a0, maxSize=5, offset=0, orderBy=name, order=asc, attributeSelect=salutationName,firstName,lastName,middleName,name,accountRole,emailAddressIsOptedOut,emailAddressIsInvalid,emailAddress,emailAddressData	list[0].emailAddress	ada.finch@example.test
qa-12-get_lead_by_uid	get_lead_by_uid	uid=6adb962fbbec3c278	name	Odile Marchbank

Writes

Graded from the environment's state through data/get (scores.state in the episodes table; reward 1 only if every check passes).

task	prompt	checks
write-update	In this EspoCRM, rename the task "Measure the Blue Heron cold room" to "Measure the Blue Heron cold room and pantry", set its status to Completed, and set its description to "Measured: north wall 6.4 m, pantry 2.1 m."	the task has the new name; its status is Completed; it has the new description; no task was added or removed
write-update-lookup	In this EspoCRM, one task is waiting for the lamp workshop lease to be signed. The lease is now signed: set that task's status to Started and its description to "Lease signed; walk row D with the lamp workshop crew.", keeping its name.	the row D floor plan task is Started; it has the new description; another task is unchanged; no task was added or removed

Episodes

model	retrieval episodes	passed	pass rate	mean tool calls	mean input tokens
openai/gpt-oss-120b:groq	12	12	100%	1.0	52,952
zai-org/GLM-5.3-Flash:zai-org	12	12	100%	1.0	66,844

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

model	write episodes	passed
openai/gpt-oss-120b:groq	2	2
zai-org/GLM-5.3-Flash:zai-org	2	2

The task create and delete are left out rather than faked. EspoCRM derives a created task's dateEnd (the day after dateEndDate) and its accountName and parentName from the sent date and parent id; REA has no rule for either derivation, and the two recorded creates send different dates and parents. A delete is a soft delete: the task stays readable by id with deleted: true, which REA does not model. The task edit is served: it echoes the name, status and description it is sent, stamps the acting user as the modifier, advances modifiedAt and streamUpdatedAt, and increments versionNumber. It accepts only edits that send all three members, as every recorded edit did.

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/EspoCRMEnv-REA --bundle espocrm --task smoke          # Docker, no model
agent-env hf run earakely-scale/EspoCRMEnv-REA --bundle espocrm --eval espocrm-retrieval --model openai/gpt-oss-120b:groq
agent-env hf run earakely-scale/EspoCRMEnv-REA --bundle espocrm --eval espocrm-writes --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. EspoCRM 10.0.10 with MariaDB 11.8.9 ran in Docker on a throwaway VM, bound to loopback, seeded through its own REST API with synthetic records (2 teams, 4 accounts, 4 contacts, 3 leads, 4 opportunities, 3 cases, 3 tasks, 2 meetings, 2 calls, 2 knowledge-base articles, a campaign, a target list and a stream post); people and places are invented.
2. Playwright drove a login, the main list and detail views and a scripted read sequence through the app's JSON API, then created, edited and deleted two scratch tasks and edited a seeded task, reading each write back at once by item and by the task list; the HAR was scoped to the API, the login exchange and the web client's UI-string and metadata calls dropped, secret-looking string 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; the write tasks were written by hand against the served write tools and the seeded names; 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	EspoCRM (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)	86 tools mirroring the application's endpoints
State	Full: a virtual clock, triggers, frozen windows, agents that answer	A state extracted from the capture (129 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, state checks for writes; 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 EspoCRM; 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 EspoCRM instance over synthetic data; EspoCRM is AGPL-3.0 software by EspoCRM, 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.