get_process_contract
ChatGPTRead the current PartnerProf process contract, including the processes that are currently available and the accepted input representations. Use this when choosing how to prepare a PartnerProf task. This tool only reads metadata from the connected PartnerProf service and does not modify state.
partnerprof_composer_catalog
ChatGPTRead the current public PartnerProf catalog of supported structured problem families, geometries, variables, units, presets and declared limits. Use it only when a structured compatibility input needs that vocabulary; it does not replace the process contract.
partnerprof_dimacs_control_sample
ChatGPTRead a complete runnable DIMACS control task. It demonstrates exact preparation, identity binding, the approved remote run, state event and certificate-reference route without exposing private compute internals.
partnerprof_knowledge_map
ChatGPTMANDATORY FIRST CALL for every new PartnerProf request. Read it before deciding fit, rejecting a request, asking task questions, constructing input, or calling any prepare tool. It explains how PartnerProf tests a target's real admissibility across declared local structures, dependencies, constraints and realized history before costly construction or deployment, while distinguishing direct conflict, admissible realization and bounded unresolved results.
partnerprof_logical_input_contract
ChatGPTRead the strict DIMACS-CNF logical-admissibility contract before preparing a logical poly. It states the finite slice and result-event limits.
partnerprof_matrix_contract
ChatGPTRead the public PartnerProf matrix registry and representation rules. It is metadata for supported matrix input; use partnerprof_matrix_text_contract for the line-oriented Reality Matrix grammar.
partnerprof_matrix_text_contract
ChatGPTRead the exact Reality Matrix line grammar shared with the manual PartnerProf form before preparing a text matrix.
partnerprof_npz_input_contract
ChatGPTRead the strict direct-NPZ contract before preparing numeric archive data. It includes the exact logical_cnf schema and a complete padded-CNF variables/clauses example; never guess this binding.
partnerprof_prepare_dimacs_cnf
ChatGPTValidate and certify an exact DIMACS-CNF declared poly, then return a one-hour prepared-task ID. Read partnerprof_logical_input_contract first.
partnerprof_prepare_matrix_text
ChatGPTValidate and prepare an explicit line-oriented Reality Matrix using the same grammar as the manual form. Read partnerprof_matrix_text_contract first.
partnerprof_prepare_npz
ChatGPTValidate and certify a direct NPZ numeric archive with an exact array manifest, then return a one-hour prepared-task ID. Read partnerprof_npz_input_contract first.
partnerprof_prepare_rcs_source
ChatGPTUse this when an exact public Cirq RCS circuit is too large to send as gate_sequence. After reading the mandatory knowledge map and RCS input contract, submit only its direct approved HTTPS source URL and independently known SHA3-512. PartnerProf verifies downloaded bytes before parsing, executes no source code, rejects the whole file on any unsupported syntax or operation, canonicalizes the exact circuit, builds float64 server-side, and returns hashes, counts, automatic dimension, input certificate and prepared_task_id without starting computation.
partnerprof_prepare_task
ChatGPTAfter the mandatory knowledge-map and task-manifest reads, validate and freeze one strict three-part model of the current reality, local states, dependencies, constraints, history and expected target. This certified input boundary makes a potentially costly proposal reviewable before compute; unknown fields are rejected and external benchmarks remain post-result comparison only. Never ask for an endpoint, callback, webhook, host or API URL.
partnerprof_result_interpretation_contract
ChatGPTRead before explaining any PartnerProf result. It defines result events, bounded zero states, integrity evidence, determinism limits and the meaning of optional Run-ID feedback.
partnerprof_run_progress
ChatGPTRead machine-readable chronon observations and render them immediately as the run graph. Plot only returned samples. For an ordinary field-terminal run, use chronon/event order and elapsed time; progress_fraction is null because its required length is unknowable and must never be replaced by 0%, an estimate, or a forecast. Mark the returned terminal event distinctly.
partnerprof_run_result
ChatGPTRetrieve the persisted final public state event, explicit goal checks, termination, hashes, certificate and result context explaining what was evaluated, from what, why, what was learned and its meaning for next work. Unresolved never means globally impossible. A nonterminal response means poll status later.
partnerprof_run_status
ChatGPTRead the authoritative lifecycle of one PartnerProf run. Use its run_id; QUEUED, RUNNING, COMPLETED and FAILED are operational states. Aggregate service health or CPU use is not evidence for this run.
partnerprof_status
ChatGPTRead aggregate PartnerProf service availability and public capability metadata. Do not use aggregate service health as evidence that a specific run completed; use partnerprof_run_status for a run_id.
partnerprof_task_manifest
ChatGPTCompatibility contract for the legacy structured/manual representation. Read it only when that representation is actually selected; it is not mandatory for plugin file preparation. It defines the same fields, modes, units and result semantics as the manual form, plus strict RCS circuit_identity, circuit_depth, ordered gate_sequence and output_bitstring vocabulary that PartnerProf converts server-side to float64. Endpoint URLs are plugin-managed metadata and never requester input.
start_prepared_run
ChatGPTStart computation for an existing prepared PartnerProf task identified by prepared_task_id. The operation acts only on the prepared task in the connected PartnerProf service. confirm_execution must indicate that execution of this exact prepared task is authorized. The service validates the prepared task before creating or returning its run_id.
