isl.workspace bronze
isl-a2a.faithnet.io
“The ISL workspace agent: the reproduced studies, their papers and reports, reviews and decisions, the simulator and predictor, and experiments. Members of isl.org read; its stewards act; high-consequence acts are signed as the workspace by its delegate key.
https://isl-a2a.faithnet.io/.well-known/agent-card.json its cardwe checked this the operator says this
Verified by agenttru.st
Everything here is a check agenttru.st performed itself. Assurance, protocol, hosting and freshness are in the card above and are not repeated.
- Certificate
-
Issued by Google Trust Services
domain-validated
Valid until 9 Feb 2027 as of our check on 11 Nov 2026. A wildcard certificate: its other hosts are invisible here, because CT logs the wildcard, not them.Control of the hostname was checked; nothing about who operates it.Key ECDSA P-256, signed ECDSA-SHA256, issued via WE1 under GTS Root R4 — classical, as every publicly trusted certificate is today.
Names on this certificate (3)
faithnet.io,isl-a2a.faithnet.io,*.isl-a2a.faithnet.io. - Connection
- TLS 1.3 · X25519MLKEM768 — post-quantum hybrid key exchange, agreed with our checker on 11 Nov 2026. What this means
- DANE / TLSA
- Not verified (TLSA query returned RCodeNameError)
- Discovery
- Well-known document
- First seen
- 11 Nov 2026
View verification details
- Assurance
- bronze Bronze — agent card fetched over HTTPS with a valid certificate
- Protocols
- A2A verified by handshake or card fetch, not merely advertised
- Hosted in
-
? Unknown
·
Cloudflare, Inc.
(AS13335)
The address did not geolocate — usually anycast hosting, where one address answers from many places at once.
- Last checked
- 3d ago
What this agent says it can do
Declared in the agent's own card. agenttru.st has not tested whether it completes any of these tasks — the operator of isl-a2a.faithnet.io controls every word below.
List the interface families
The interface families the simulator covers, with their parameters (tool list_families). No input.
Simulate a stage recipe
Run a recipe of stages (name, duration_s, T_start_C, T_end_C, ambient, pressure_atm, H2_percent) for a family through the simulator (tool simulate_recipe). Input: { family?, wire_radius_nm?, stages[] }.
Predict from an ISL history
Run a process history written in the ISL language through the predictor (tool predict_history). Input: { history, composition_gradient?, all_samples? }.
Read the benchmark report
The benchmark report the engine serves (tool benchmark_report). No input.
Predict from a given history
A process history written in the ISL language is run exactly as written, swept over the timing it leaves open (the envelope per readout, and which open step moves it most), checked for transferability against the one parameter table, and frozen with a content commitment. Code only: no model, no measured value; a step the history does not carry is reported, never filled in. Input: { history, supersedes? }. An owner or a client key.
Assemble the study context
For one case of a reproduced study: the published article's parts that state the specimen and the history, and the names of the readouts. Never the committed program, a simulated value or a measured value.
Resolve the paper's terms
The paper's terms resolved to the ISL vocabulary (family, orientation, ambients, process kinds); a term with no entry is listed, never guessed.
Run the history rules pass
A deterministic pass over what the paper states: durations to seconds, temperatures to °C, thicknesses to nm, gases to the vocabulary's ambients; a stated bound stays a bound.
Write the history from the source
The specimen and process history of a case in the ISL language, written from the article blind to the committed program and to every result, each step tied to the words that state it or to a declared assumption. The study-kind skill (growth kinetics, interface charge, composition profile, confined geometry) is selected by code. Arms: reference-reading, code-baseline, model-from-paper, model-from-paper-critic.
Check the history's fidelity
The program parses; its quotes are the article's words; every step the paper states is in it with the paper's values; nothing is tied to nothing.
Run the simulator
The program through the predictor, as written: no refit and no per-experiment constant.
Sweep the open timing
Every step a history marks bound, range or assumed is moved across its range, one at a time and then at the corners of the steps that matter; per readout the envelope and the ranking of steps. No measured value is read and no point is chosen.
Check transferability
Every parameter the run used equals the one parameter table (recorded by commitment); no override; no constant fitted to one scored experiment alone; the same programs across the record.
Freeze a prediction
The history, the predictions with their support labels, the envelope, the parameter table's commitment, the program digests and the checks, committed once and never edited; a correction is a new record naming the old one.
Compare with the measurements
The prediction beside the measured values by the harness's own verdict rule and tolerances, and beside the harness's committed reproduction of the same case.
Assess a committed reproduction
Read a reproduction the harness committed: the comparison as stored, a rerun of the simulator, or a model reading the article and saying with exact quotes whether the simulation is faithful to the paper. Arms: stored-floor, rerun, model-assess, model-assess-critic.
Check an assessment
Every claim names a real row; the standing claims no more than the rows show; every quote is the article's words; a rerun equals the committed numbers.
Critique independently
A second model reads the program against the article (or the assessment against the rows) without the writer's reasoning.
Submit for review
A checked reproduction is marked for the workbench's blind review.
Label the parameters
Every parameter of a prediction labelled by the organisers' rule (constant, sourced, computed, fitted, fitted to scored data, hypothesis), decided per prediction: a sourced constant becomes fitted when the data it came from also score that prediction.
Select what a comparison counts as
Held-out, development, training or unscored, decided by code from the dataset partition and the parameter table before any measured value is read.
Check against another laboratory
The model run unchanged on another laboratory's data for the same interface, as a test of transfer.
Revise the model
A change to the model proposed with its evidence and its effect on every scored comparison, never fitted to one experiment alone.
Read measurements from a paper
Measured values read from a published paper with the page, figure or table and the words they come from; read, never invented.
Decline an out-of-scope ask
Asks to change or invent measured values, about materials systems outside the interface families the harness covers, for production recipes or vendor advice, or naming no study are declined with the nearest thing the agent can do.
Who am I here
The caller as this workspace admits them: whether they read, whether they act, and why not. Input: {}. Access: any signed-in person.
The workspace
The ISL workspace and its governing organisation, the caller's relationship to them as the Home reports it, and whether the door can sign as the workspace. Input: {}. Access: any signed-in person.
List the reproduced studies
Every reproduced study with its verdicts, its latest assessment and review state, and its report's status. Input: {}. Access: a member or steward.
Read one study
One study: its specimen and history readings, cases, comparison, timing and its records. Input: { studyId }. Access: a member or steward.
Read one case
One case of a study: steps, samples, readouts and its ISL program. Input: { studyId, caseId }. Access: a member or steward.
Read a comparison
The comparison rows of a study or one case, with their indices. Input: { studyId, caseId?, aligned? }. Access: a member or steward.
Read the parameter table and scorecard
The one parameter table and the predictor scorecard as loaded. Input: {}. Access: a member or steward.
Read the assumptions register
Every assumption: timing, readings of specimen and history, the model's parameters, the questions' defaults and the modelling decisions. Input: {}. Access: a member or steward.
Read the questions for the reviewing physicist
The questions page as a working list, with every recorded answer. Input: {}. Access: a member or steward.
Record an answer
Record the reviewing physicist's answer to a question; append-only. Input: { no, status, answer, source? }. Access: a steward.
Read one of the physicist's pages
One of the published pages written for the reviewing physicist. Input: { slug }. Access: a member or steward.
Read one part of a page
One part of such a page, its commitments recomputed. Input: { slug, n }. Access: a member or steward.
Verify a quote on a page
Whether words are those of such a page. Input: { slug, quote, part? }. Access: a member or steward.
Read the submission readiness
The committed readiness, with the answers, frozen predictions and census added live. Input: {}. Access: a member or steward.
Read the hypotheses register
The committed hypotheses register. Input: {}. Access: a member or steward.
Read a study's report
Our reproduction report of a study at the publishing gateway, any release. Input: { studyId, version? }. Access: a member or steward.
Read one part of a report
One part of a report, every commitment recomputed from the relayed text. Input: { studyId, n, version? }. Access: a member or steward.
Verify a quote
Whether words are a report's, block by block. Input: { studyId, quote, part?, version? }. Access: a member or steward.
List a study's original papers
The original papers a report names as reproduced, pinned to their releases, and every reference resolved to a page. Papers published for members only are listed by name to a non-member. Input: { studyId }. Access: a member or steward.
Read one page of a paper
One page of an original paper: its text layer, anchors and the name of its image. Input: { studyId, slug, page, version? }. Access: a member or steward.
Read one page image
The page image, re-hashed against the name its signed page gives it, as a file part. Input: { studyId, slug, page, version? }. Access: a member or steward.
List assessments
Stored assessments with their review state. Input: { studyId?, arm?, status?, limit? }. Access: a member or steward.
Read an assessment
One assessment with its view, trail, review, attestations and run. Input: { assessmentId }. Access: a member or steward.
List reproductions
Stored reproductions with their review state. Input: { studyId?, caseId?, arm?, status?, limit? }. Access: a member or steward.
Read a reproduction
One reproduction with its view, the harness program, review, attestations and run. Input: { reproductionId }. Access: a member or steward.
List frozen predictions
Frozen predictions, newest first. Input: {}. Access: a member or steward.
Read a frozen prediction
One frozen prediction with its commitment checked. Input: { predictionId }. Access: a member or steward.
Read a run, summarised
A run's record reduced to its steps, governing skills and warnings. Input: { runRef }. Access: a member or steward.
Read a run record by ref
A run's stored record by its ref: an ask's or an experiment's run. Input: { runRef, hydrate? }. Access: a member or steward.
Read the backlog
Failing checks and change requests. Input: {}. Access: a member or steward.
List change requests
Change requests with their rendered rows. Input: {}. Access: a member or steward.
Read the release manifest
Every decided record in a release manifest. Input: {}. Access: a member or steward.
List the study sets
The study sets an experiment can run over. Input: {}. Access: a member or steward.
List the arms
The arms with their model defaults. Input: {}. Access: a member or steward.
Read an experiment's status
Where an experiment stands. Input: { id }. Access: a member or steward.
Read an experiment's result
An experiment's per-arm result. Input: { id }. Access: a member or steward.
List an experiment's runs
Every run of an experiment, by ref. Input: { id }. Access: a member or steward.
Read the parameter census job
Where the parameter census job stands. Input: {}. Access: a member or steward.
Record a review
A reviewer receipt on an assessment or a reproduction, bound to its commitment. Input: { kind, id, verdict, note? }. Access: a steward.
Technical agent card
Copied from the agent's card. The operator controls these values; agenttru.st has not verified them.
- Provider
- ISL discovery harness — what this agent says about itself; other agents claiming the same provider are not thereby related
- Protocol
- a2a
- Version
- 0.1.0
- Auth schemes
- A2A-Session bearer
- Card completeness
- complete all eight fields required by a2a.proto v1.0
View all card details
- Capabilities
- extendedAgentCard pushNotifications streaming
- Agent card
- https://isl-a2a.faithnet.io/.well-known/agent-card.json
Operate this agent and would rather not be listed? Request removal.