Search Space¶

The harness research pipeline explores a space of 263,424 possible configurations.

Dimensions¶

8 × 7 × 7 × 7 × 2 × 2 × 3 × 4 × 2 = 263,424

Axis	Values	Count
`max_iterations`	3, 4, 5, 6, 7, 8, 9, 10	8
`thinking_budget`	256, 512, 1024, 2048, 3072, 4096, 8192	7
`context_budget`	2048, 4096, 8192, 16384, 24576, 32768, 65536	7
`cost_gate_usd`	0.001, 0.005, 0.01, 0.02, 0.05, 0.10, 0.25	7
`retry_on_error`	true, false	2
`drift_detection`	true, false	2
`p66_consent_threshold`	physical, digital, none	3
`pattern`	single_agent_supervisor, initializer_executor, multi_agent, reactive	4
`memory_backend`	working, filesystem	2

Progress¶

Metric	Value
Total configs	263,424
Explored	435 (0.17%)
Champion score	0.6541
Est. full exploration	~1,800 robot-days

Named candidates¶

The pipeline includes 56 named "interesting" candidates with human-readable IDs:

lower_cost — minimal cost gate, conservative budgets (current champion)
max_thinking — maximum thinking budget, high context
fast_low_cost — low iterations, minimal cost
balanced_retry — moderate everything, retry enabled
no_drift — drift detection disabled for comparison
... and 51 more

State tracking¶

The search space state is tracked in:

opencastor-ops/harness-research/search_space_state.json

{
  "explored_count": 435,
  "last_updated": "2026-03-21T...",
  "champion_id": "lower_cost",
  "champion_score": 0.6541
}

API¶

GET /api/research/status

{
  "search_space_size": 263424,
  "explored_count": 435,
  "explored_pct": 0.17,
  "champion": {
    "candidate_id": "lower_cost",
    "score": 0.6541
  }
}