Thore Cimbal
33a0ee7aec
Three minor releases at once. Byte-identical surface: only WORKFLOW.md actually changed (+128/-2); CLAUDE.md and the four templates are untouched. AGENTS.md takes the new upstream prefix (+22/-2) and keeps our project section unchanged. The two declared-extended files were reconciled by hand, because nothing compares them: validate.py gains upstream's check_vendored_portable, and schema.yaml gains vendored, project_section_marker, the judged link field and the ledger and verdict types. gen_status.py needed nothing — upstream did not touch it. Recorded in the new HERKUNFT.md, including that this reconciliation has no contradictor and will be forgotten next time. Newly adopted: check_harvest.py and judge.py, plus docs/ledger/ and docs/verdict/ with their templates. Deliberately not adopted: check_locked.py — pruefe_sperrliste.py has done that job here since 2026-08-20, and two tools for one rule is maintenance without gain. The v0.1.1 baseline stays where it is; docs/sources is immutable.
61 lines
2.0 KiB
Markdown
61 lines
2.0 KiB
Markdown
---
|
|
type: verdict
|
|
date: YYYY-MM-DD
|
|
outcome: clean # clean | model-failure | framework-gap | both
|
|
judged: docs/ledger/YYYY-MM-DD-slug.md
|
|
related: []
|
|
---
|
|
|
|
<!-- Copy to docs/verdict/YYYY-MM-DD-slug.md. Delete all comments when filling in. -->
|
|
|
|
# Verdict: <the run this judges>
|
|
|
|
⚠️ **Written in fresh context, without the work transcript.** An agent that
|
|
judges its own session justifies rather than checks. The judge receives
|
|
`AGENTS.md`, `WORKFLOW.md`, the diff and the ledger — nothing else. A
|
|
verdict produced in the working session is void, whatever it says.
|
|
|
|
## Deterministic pass
|
|
|
|
<!-- Quote the run of scripts/judge.py. Its findings are facts about the
|
|
trace; everything below is inference on top of them. -->
|
|
|
|
```
|
|
judge: …
|
|
```
|
|
|
|
## Rubric
|
|
|
|
<!-- Five questions, each answered with a quotation from the diff or the
|
|
ledger rather than an impression. "Looks fine" is not an answer. -->
|
|
|
|
**1. Scope.** Does every changed file trace to the stated undertaking?
|
|
|
|
**2. Substance.** Is the design document load-bearing — do the non-goals
|
|
exclude something a reader would otherwise expect, are the shakiest calls
|
|
real risks?
|
|
|
|
**3. Size.** Was the declared class plausible for what the diff became?
|
|
|
|
**4. Evidence.** Where a slice claims a result, is a run quoted? Where a
|
|
check was added, was it shown going red?
|
|
|
|
**5. Silence.** Which rule should have left a trace here and did not?
|
|
|
|
## Findings
|
|
|
|
<!-- Every finding lands in exactly one bucket. Getting this wrong is worse
|
|
than missing the finding: the wrong bucket either blames a person for a
|
|
rule that does not exist, or writes a behavioural lapse into the rule
|
|
set as though it were a design defect. See ADR-0010. -->
|
|
|
|
| bucket | rule | finding |
|
|
|---|---|---|
|
|
| model-failure | | a clear rule was not followed |
|
|
| framework-gap | | the rule is missing, unenforceable, or invisible |
|
|
|
|
## What follows
|
|
|
|
<!-- Framework gaps become issues. Model failures do not — they are reported
|
|
and left as behaviour. Say which is which and why. -->
|