"SCIENTIA A2A evidence-gathering tasks"

SCIENTIA A2A evidence-gathering tasks

Orchestrator / mesh A2A can delegate read-heavy, idempotent jobs that return structured JSON for metadata_json.scientia_evidence or publication_status_events. This document names task kinds for operators and agent authors; routing uses existing RemoteTaskEnvelope types in vox-orchestrator (a2a / envelope modules).

Allowed task families

Task kind (logical)GoalMust not
scientia.gather.benchmark_lineageCollect baseline/candidate run ids and report pathsInvent benchmark outcomes
scientia.gather.repo_docsList ADR/research paths and linked corpusSummarize novelty
scientia.gather.repro_artifactsFind checksum / manifest pathsClaim reproducibility passed
scientia.gather.venue_requirementsFetch venue checklist text (cached)Assert submission eligibility
scientia.gather.credential_presenceClavis/env presence bits onlyExpose secret values

Envelope rules

  1. Payload is JSON with task_kind, publication_id, repository_id (when known), and idempotency_key.
  2. Result merges into scientia_evidence or appends a status event with detail_json pointing at file paths and digests.
  3. Refusal: if grounding artifacts are missing, return blocked_reasons — never backfill with LLM prose.
  4. Human loop: meaningful advance, novelty, and final abstract remain human-attested per how-to: Scientia publication.
  • Discovery ranking: vox_scientia_publication_discovery_scan / vox scientia publication-discovery-scan
  • LLM assist (bounded): vox_scientia_assist_suggestions (use_llm=false for heuristic-only)