Worked process/campaign document examples + metric-vocabulary enumeration #197

New Issue

2026-07-03T16:52:11+02:00

Brummel commented

2026-07-03 16:52:11 +02:00

First-time authoring has no worked reference (fieldtest cycle-0106, F6): no process/campaign JSON example exists in docs/ or the corpus (the cycle-0106 fieldtest corpus now carries the first ones), block hints say 'metric name (see metric_vocabulary)' but no verb lists that vocabulary, and the 'unknown metric' refusal does not enumerate valid names (they are only discoverable via the glossary R-metrics entry).

Likely shape: one worked example per document type in user-facing docs; metric enumeration via 'aura process introspect --metrics' (or listing valid names in the unknown-metric refusal).

depends on: nothing
context: fieldtest finding F6, fieldtests/cycle-0106-research-artifacts/FINDINGS.md (2026-07-03); the docs half is docwriter territory post-stability.

First-time authoring has no worked reference (fieldtest cycle-0106, F6): no process/campaign JSON example exists in docs/ or the corpus (the cycle-0106 fieldtest corpus now carries the first ones), block hints say 'metric name (see metric_vocabulary)' but no verb lists that vocabulary, and the 'unknown metric' refusal does not enumerate valid names (they are only discoverable via the glossary R-metrics entry). Likely shape: one worked example per document type in user-facing docs; metric enumeration via 'aura process introspect --metrics' (or listing valid names in the unknown-metric refusal). - depends on: nothing - context: fieldtest finding F6, fieldtests/cycle-0106-research-artifacts/FINDINGS.md (2026-07-03); the docs half is docwriter territory post-stability.

Brummel commented

2026-07-03 22:28:56 +02:00

Re-confirmed by the cycle-0107 fieldtest (F9) — the roster split is now user-visible

The executor cycle sharpened this gap: there are now TWO rosters a campaign author must know and neither is enumerable from the public surface. process validate accepts any metric string intrinsically only if it is in the 17-name metric_vocabulary — but the executor preflight then refuses selection metrics outside the registry's 7-name rankable set (process stage 0: metric "profit_factor" is not rankable (winner selection needs one of the registry's rankable metrics)) and gate metrics outside the 14-name per-member set. profit_factor is emitted in every member's metrics block yet is not rankable, so the exists-but-not-rankable split is invisible without trial and error. Suggested shape (unchanged from the body, now with the split): expose the rosters via introspection (e.g. process introspect --metrics listing the 17 names annotated with rankable / per-member applicability) or enumerate the applicable roster in each refusal.

## Re-confirmed by the cycle-0107 fieldtest (F9) — the roster split is now user-visible The executor cycle sharpened this gap: there are now TWO rosters a campaign author must know and neither is enumerable from the public surface. `process validate` accepts any metric string intrinsically only if it is in the 17-name `metric_vocabulary` — but the executor preflight then refuses selection metrics outside the registry's 7-name rankable set (`process stage 0: metric "profit_factor" is not rankable (winner selection needs one of the registry's rankable metrics)`) and gate metrics outside the 14-name per-member set. `profit_factor` is emitted in every member's metrics block yet is not rankable, so the exists-but-not-rankable split is invisible without trial and error. Suggested shape (unchanged from the body, now with the split): expose the rosters via introspection (e.g. `process introspect --metrics` listing the 17 names annotated with rankable / per-member applicability) or enumerate the applicable roster in each refusal.

Brummel referenced this issue

2026-07-04 04:45:36 +02:00

Document the op-script grammar on the public surface (graph introspect / project-layout) #208

Brummel referenced this issue from a commit

2026-07-04 10:53:53 +02:00

fieldtest: cycle-0107 campaign executor + on-ramp corpus (GREEN-core)

Brummel referenced this issue from a commit

2026-07-04 10:53:54 +02:00

feat(cli): process|campaign introspect --metrics — the roster split becomes discoverable (0107 F9)

Brummel referenced this issue from a commit

2026-07-04 10:53:54 +02:00

fieldtest: milestone-research-artifacts corpus (GREEN) + Aura.toml data-key doc fix (F7)