git:20261003.82faac0 to git:20261003.bfd7d97
4 added, 3 removed. Audit A to A.
# evidence Agent Contract
Frozen artifacts for round 158: the validation record, the replay runners
behind its numbers and their per-run record, and the review pass records with
their dispositions.
Rules:
- **Frozen after landing.** Do not edit, regrade or regenerate these files to
improve a later conclusion. Overturn a reading with a new measurement in a new
round's evidence directory and cite it from that round's register row.
- **The runners are evidence tools, not repository gates.** They call live
models, are non-deterministic, and must never be wired into `make test` or any
blocking check. A run proves only what one model produced on one prompt.
- **Read the runner before citing a number.** Each one disables tools and hooks,
- runs in an empty directory and states in its docstring what it counts; the
- regex graders are recall aids, and the validation record says which readings
- were checked by eye.
+ runs in an empty directory and states in its docstring what it counts. The
+ regex graders are recall aids: the scope counts in the validation record come
+ from by-hand readings kept in the per-run records, and the triage replay reads
+ a marker the model emits.
- **Neutral domain only.** Scenarios use a synthetic report-export service; no
credentials, host paths, source sessions' product terms or requester text.
Validation:
- `python3 -m py_compile specs/158-review-finding-scope/evidence/*.py`
- `bash skills/skill-extraction-workflow/scripts/check-ccl-skills.sh .`