cyberneticlibrary
← library

Tag and discover benchmark tasks

enrich-task-tagsworkflowsetup L2★12
SShadowS/CentralGauge ↗
What it does

fan out and consolidate task tags

Best for

Batch tagging homogeneous content (e.g., all AL benchmark tasks) against controlled vocabulary.

Inputs
  • · {codexOutput?, opencodeOutput?, priorRoundFindingCount?, roundNum?}
  • · [{file, count, diagPath}]
  • · [{id, name, spec}]
Outputs
  • · Analysis report with recommendations
  • · Drafted specification document
  • · Tagged items with controlled facets
Requires
  • · Read tool
  • · HTTP/API
  • · Claude API (subagent)
Preconditions
  • · Benchmark task specs available
Failure modes
  • · Schema validation failure (agent output mismatch)
  • · Subagent timeout or rejection
  • · Unable to parse freeform output (parse phase failure)
  • · Agent cannot access required files
Trust signals
  • · JSON Schema validation enforced on every output
  • · Cross-reviewer deduplication implemented