NW-RIGOR
The user collaborates with you. You run every command; the user never types
terminal commands. nwdes does not execute models, and the CLI neither maps
competence to a model, checks availability, nor falls back.
1. Read what is already set
Run nwave-ai rigor --show (read only) and read its values and their sources.
Also read the project .nwave/config.json and the global .nwave/config.json
(under a valid absolute NWAVE_AGENTS_HOME, otherwise the home directory).
Reuse anything the user already said in this conversation; never ask them to
repeat a known value or type an identifier already available.
2. Ask once, in one batch
Ask ONE batched question set in plain words (short sentences, everyday words),
not one round trip per item. Say what each item changes for the user's work.
Keep tool names (option names, producers, reviewers, executors, role
names, charter, slice, lane, handoff) in your commands only, never in
a sentence the user reads. Cover only what is missing or what the user wants
to change. Each item gives: the current value and source as the suggested
answer, the cost and benefit of the options, and your recommendation. End with
one direct question ("keep these, or change which?").
Review rounds (1-5). How many times a second person, not the author,
may check the same piece of work. Cost: one more check per round. Benefit:
fewer mistakes stay unseen. Too few rounds can miss mistakes; many rounds
can keep finding small things and never stop.
Examine level. A final check of the finished work, by someone who has
not read the code; it finds what automatic tests miss. Explain "examination"
as one such check and "repair round" as one chance to fix what it found.
Present the tool's own text, copied here once and pinned by a test:
| Level |
gives |
not covered at this level |
quick |
1 examination, no repair round |
any repair round |
standard (default) |
2 examinations, 1 repair round |
more than 1 repair round |
extended |
3 examinations, 2 repair rounds |
more than 2 repair rounds |
Not covered at any level: security, load or checks on other machines.
The tool also says: a row not examined is never green. A longer level is not safer.
Which AI model does which kind of work. Say it as: a stronger model for
hard work, a faster one for routine work, for all work of one kind or for one
named helper. Offer the model already in effect or known from the config or
the conversation; keeping it is the default. Map the answer to
--model-set with a group (producers, reviewers, executors) or an
exact published role name.
Readable summary pages. A web page in plain language, for people who do
not read code, written when a piece of work is finished: yes or no. When
yes, also the language (a code such as en, de, it).
Other AI tools on this computer. Whether they should also get the model
choices. Writing changes files in the user's home folder; a backup copy
is kept next to each. Two tools (codex, devin) cannot take a model per
kind of work and refuse. Default: change nothing.
Rounds and levels bound the rework loops. A finding still open at the limit
returns to its owner; reaching the limit is never an approval and nothing
passes silently.
3. Apply with ONE command
The command takes every chosen value together. State it before running, run it,
report the result, then run nwave-ai rigor --show and show the new values.
Use current or keep for what does not change, and --yes: the user's
answers are the confirmation. Give each option once.
nwave-ai rigor --yes --review-rounds 3 --examine-level extended --models keep --documents yes --document-language en --hosts keep
That example sets 3 rounds, the extended level and English documents, and
changes no models or host file. Adapt only the values. The options:
--show: print the values in effect; writes nothing.
--yes: the user's answers are the confirmation.
--review-rounds: 1 to 5, or current.
--examine-level: quick, standard, extended, or current.
--models: keep; or use --model-set once per target, as
--model-set reviewers=PROVIDER/MODEL.
--documents: yes, no or keep; --document-language LANG only with yes.
--hosts: host names separated by commas, none or keep.
After a refusal, report its WHAT/WHY/HOW and ask the user; never substitute
another value, provider, model or scope.
Provider/model selection
Model facts for native-host role invocation are set per target: --model-set
above (groups producers, reviewers, executors, or a published role name)
or, for the default or one role, nwave-ai model set:
| Scope and target |
Command |
| Global default |
nwave-ai model set --provider PROVIDER --model MODEL |
| Current-project default |
nwave-ai model set --provider PROVIDER --model MODEL --project |
| Global published-role override |
nwave-ai model set --provider PROVIDER --model MODEL --role ROLE |
| Current-project published-role override |
nwave-ai model set --provider PROVIDER --model MODEL --role ROLE --project |
- Suggest the effective current pair (project role override beats global
default) with its source; the user accepts or changes it.
- A requested competence level helps the user choose a pair; the CLI has no
competence profile or mapping. Keep the current pair until the user selects
another.
PROVIDER is codex or claude; MODEL the accepted exact string; ROLE
only an exact published role name. Never invent an identifier, a competence
mapping or a fallback; the stored value is not checked, so a wrong name shows
only when the role runs.
1---2name: nw-rigor3description: Use when the user wants to see or change how much rigor a project gets (review rounds, examine level), which provider/model a role uses, or whether readable documents are written. The LLM reads the current settings, asks only the missing choices in plain words, and runs `nwave-ai rigor` or `nwave-ai model set` itself. Never derives a model from a competence label.4user-invocable: true5argument-hint: '[review rounds, examine level, provider, model, scope, optional role]'6metadata:7 source: https://skillmd.com/skills/nwave-ai/nw-rigor8---910# NW-RIGOR1112The user collaborates with you. You run every command; the user never types13terminal commands. nwdes does not execute models, and the CLI neither maps14competence to a model, checks availability, nor falls back.1516## 1. Read what is already set1718Run `nwave-ai rigor --show` (read only) and read its values and their sources.19Also read the project `.nwave/config.json` and the global `.nwave/config.json`20(under a valid absolute `NWAVE_AGENTS_HOME`, otherwise the home directory).21Reuse anything the user already said in this conversation; never ask them to22repeat a known value or type an identifier already available.2324## 2. Ask once, in one batch2526Ask ONE batched question set in plain words (short sentences, everyday words),27not one round trip per item. Say what each item changes for the user's work.28Keep tool names (option names, `producers`, `reviewers`, `executors`, role29names, `charter`, `slice`, `lane`, `handoff`) in your commands only, never in30a sentence the user reads. Cover only what is missing or what the user wants31to change. Each item gives: the current value and source as the suggested32answer, the cost and benefit of the options, and your recommendation. End with33one direct question ("keep these, or change which?").34351. **Review rounds (1-5).** How many times a second person, not the author,36 may check the same piece of work. Cost: one more check per round. Benefit:37 fewer mistakes stay unseen. Too few rounds can miss mistakes; many rounds38 can keep finding small things and never stop.392. **Examine level.** A final check of the finished work, by someone who has40 not read the code; it finds what automatic tests miss. Explain "examination"41 as one such check and "repair round" as one chance to fix what it found.42 Present the tool's own text, copied here once and pinned by a test:4344 | Level | gives | not covered at this level |45 |---|---|---|46 | `quick` | 1 examination, no repair round | any repair round |47 | `standard` (default) | 2 examinations, 1 repair round | more than 1 repair round |48 | `extended` | 3 examinations, 2 repair rounds | more than 2 repair rounds |4950 Not covered at any level: security, load or checks on other machines.51 The tool also says: a row not examined is never green. A longer level is not safer.523. **Which AI model does which kind of work.** Say it as: a stronger model for53 hard work, a faster one for routine work, for all work of one kind or for one54 named helper. Offer the model already in effect or known from the config or55 the conversation; keeping it is the default. Map the answer to56 `--model-set` with a group (`producers`, `reviewers`, `executors`) or an57 exact published role name.584. **Readable summary pages.** A web page in plain language, for people who do59 not read code, written when a piece of work is finished: yes or no. When60 yes, also the language (a code such as `en`, `de`, `it`).615. **Other AI tools on this computer.** Whether they should also get the model62 choices. Writing changes files in the user's home folder; a backup copy63 is kept next to each. Two tools (`codex`, `devin`) cannot take a model per64 kind of work and refuse. Default: change nothing.6566Rounds and levels bound the rework loops. A finding still open at the limit67returns to its owner; reaching the limit is never an approval and nothing68passes silently.6970## 3. Apply with ONE command7172The command takes every chosen value together. State it before running, run it,73report the result, then run `nwave-ai rigor --show` and show the new values.74Use `current` or `keep` for what does not change, and `--yes`: the user's75answers are the confirmation. Give each option once.7677```bash78nwave-ai rigor --yes --review-rounds 3 --examine-level extended --models keep --documents yes --document-language en --hosts keep79```8081That example sets 3 rounds, the extended level and English documents, and82changes no models or host file. Adapt only the values. The options:8384- `--show`: print the values in effect; writes nothing.85- `--yes`: the user's answers are the confirmation.86- `--review-rounds`: `1` to `5`, or `current`.87- `--examine-level`: `quick`, `standard`, `extended`, or `current`.88- `--models`: `keep`; or use `--model-set` once per target, as89 `--model-set reviewers=PROVIDER/MODEL`.90- `--documents`: `yes`, `no` or `keep`; `--document-language LANG` only with `yes`.91- `--hosts`: host names separated by commas, `none` or `keep`.9293After a refusal, report its WHAT/WHY/HOW and ask the user; never substitute94another value, provider, model or scope.9596## Provider/model selection9798Model facts for native-host role invocation are set per target: `--model-set`99above (groups `producers`, `reviewers`, `executors`, or a published role name)100or, for the default or one role, `nwave-ai model set`:101102| Scope and target | Command |103|---|---|104| Global default | `nwave-ai model set --provider PROVIDER --model MODEL` |105| Current-project default | `nwave-ai model set --provider PROVIDER --model MODEL --project` |106| Global published-role override | `nwave-ai model set --provider PROVIDER --model MODEL --role ROLE` |107| Current-project published-role override | `nwave-ai model set --provider PROVIDER --model MODEL --role ROLE --project` |108109- Suggest the effective current pair (project role override beats global110 default) with its source; the user accepts or changes it.111- A requested competence level helps the user choose a pair; the CLI has no112 competence profile or mapping. Keep the current pair until the user selects113 another.114- `PROVIDER` is `codex` or `claude`; `MODEL` the accepted exact string; `ROLE`115 only an exact published role name. Never invent an identifier, a competence116 mapping or a fallback; the stored value is not checked, so a wrong name shows117 only when the role runs.