Repository navigation
domain-driven-design: glossary layout, pressure-test observe, write-for-humans eval #7591
Description
Parent
Refs #7555. Source: PLAN Phase 23, topic poteto-review. Thread: Other plugins.
Depends on
- config-cascade: settings resolution and convention home #7571 Settings resolution and convention home (glossary path settings)
Outcome
/domain-driven-design:curate-language proposes a GLOSSARY.md per bounded context plus a
GLOSSARY-MAP.md; /prototype:pressure-test gains observe; write-for-humans gets a grounding rule
only if an eval shows the gap.
Decisions
/domain-driven-design:curate-languageproposesGLOSSARY.mdper bounded context plus
GLOSSARY-MAP.md, stays domain-only; replace the fourUBIQUITOUS-LANGUAGE.mdlookups; glossary
grouping line (D70, Q51, Q75, Q525; Matt's current glossary layout). Setting: glossary paths
GLOSSARY.md,GLOSSARY-MAP.md.- Teach glossary rules; questionnaire template drops em dashes (Q525).
/prototype:pressure-testgainsobserve; prototype evidence on a throwaway branch after the
secrets check (Q167, Q168; combined: Lauren and Matt both keep prototype branches)./docs-hygiene:write-for-humansgrounding rule only if an eval shows the gap (Q526).- Rejected:
CONTEXT.md(no difference in a 4-agent experiment, "context" overloaded); a third
prototype skill; prototype evidence as a note only or pushed without the check.
Key interfaces
From the design's module layout: "domain-driven-design: curate-language proposes GLOSSARY.md
per bounded context plus GLOSSARY-MAP.md." "prototype: pressure-test gains observe."
Settings: "Glossary paths GLOSSARY.md, GLOSSARY-MAP.md."
Slices, filed when this phase starts: glossary, pressure-test-observe, write-for-humans-eval.
Done when
-
rg -c 'UBIQUITOUS-LANGUAGE' plugins --glob '!**/CHANGELOG.md'returns 0 (nonzero on main) and
bash scripts/check-purged-em-dashes.shexits 0. - The write-for-humans eval slice carries a
Probe result:comment. - Every slice sub-issue is closed by a merged PR.
Out of scope
- Dossier branches (Dossier thread), beyond sharing the secrets check.
Activity
github-actions commentedon Oct 16, 2026
Proposed work class: work-class: scoped
Basis: automated intake triage; the lanes App applied this class.
Size: Medium
Complexity: Medium
Uncertainty: Known
Flags: none
Blocked by: #7571
Understood: update curate-language in domain-driven-design to propose a GLOSSARY.md per bounded context plus a GLOSSARY-MAP.md, replacing the UBIQUITOUS-LANGUAGE.md lookups. Add
observeto prototype's pressure-test. Add a write-for-humans grounding rule only if an eval shows the gap. The outcome, decisions and done-when checks are specific, so I'm asking nothing further. The issue lists #7571 (glossary path settings) as a dependency, so I've recorded it as a blocker. I didn't find a likely duplicate among the open issues.Automated triage, applied by the lanes App.
kyle-sexton commentedon Oct 24, 2026
Released for re-triage under the owner's 2026-10-24 rule: an issue stops for a person only for an intent or vision question, a security call, a host or attended run, or spend outside prepaid credit. The hold label came off and intake triage runs again now; it re-applies a hold, with a numbered question, when one of those applies.
To hold this issue again, add
needs-human(orstatus: needs-decision) and say why in a comment.22 remaining items
kyle-sexton commentedon Nov 1, 2026
Decision recorded.
Issue Triage session (agent), applying a decision recorded in the poteto-review interview ledger; this comment asserts no new owner decision.
The issue lane stopped here with reason
intent(comment 6556369883), asking what/prototype:pressure-test observedoes and what check shows it is done. The poteto-review interview answered that as Q167. Its ledger records the owner's answer as "accepted as recommended audit A7" (.work/poteto-review/interview-checklist.md:212, a local session record), and decision row 99 records Q160, Q162 and Q165-Q169 as "APPROVED as recommended" (.work/poteto-review/GUIDE.md:209, owner 2026-10-15). The session's question the owner accepted, verbatim: "Should/prototype:pressure-testgain anobserveoption where the agent runs the variants of a fork a run can settle, records behavior, output and timing (timing labeled unqualified), and recommends, handing off to/performance:goalonly when the fork is itself a performance target?" Recommended answer: "Yes, as described; the human keeps product and preference calls; the STOP gate for unrequested prototypes stays." Q168 (evidence on a throwaway branch after the repository's secrets and machine-data check) is recorded the same way (interview-checklist.md:213).The remaining work is split into sub-issues, each with its own done-when:
observeoption for/prototype:pressure-test(Q167)./prototype:pressure-test, defined once inplugins/prototype/context/discipline.md(Q168). It records the evidence branch as an exception to rule 6 and replaces the rejection in thedocs/upstream/mattpocock-skills.mdprototype row, citing Q168. It adds no new check: it uses the repository's own secrets and machine-data check, and pushes nothing when that check is absent or fails. feat(prototype): keep explore-directions and build-generator evidence on a throwaway branch after the repository's secrets check #12030 applies it to/prototype:explore-directionsand/prototype:build-generator./planning:planlists prototype evidence and what stays unproven (Q168's planning half).This issue closes when every sub-issue closes by a merged pull request.
needs-humancomes off: the intent question has a recorded answer.github-actions commentedon Nov 3, 2026
melodic-automation-lanes commentedon Nov 3, 2026
This records a re-fire of an earlier refusal, not a new refusal. The lanes kill switch stopped run 41724819404, the work-resolve-issue run that
agent-readystarted, before it did any work; that run ended at 2026-11-03T23:45:18Z. work-unblock has now removed and addedagent-readyback to start a fresh run. A kill-switch refusal is no fault in this issue or the lanes, so it never counts against later re-fires; after limits.limit-stops-per-issue of them since a person last addedagent-readyor a run last reported, work-unblock waits longer before each one, up to 24 hours.github-actions commentedon Nov 3, 2026
github-actions commentedon Nov 4, 2026
melodic-automation-lanes commentedon Nov 4, 2026
The work-resolve-issue lane stopped on this issue (
verified-done): the issue is a finished parent, every sub-issue closed as completed, and the run checked each of its own done-when items against the default branch, citing each check, so the issue closes.It closes the issue as completed.
In the run's words:
The run's report:
What the run says it did not do:
Run: https://github.com/melodic-software/claude-code-plugins/actions/runs/41743077771