You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
List enforcement-placing skills and give each one route to the ladder - part 2 of 2 #16029
I read this as part 2 of the split of #15729. Its one done-when item is that the follow-on items for the skill-text changes exist under the container, and that their count is stated on this item. It looks like it depends on the list of enforcement-placing skills from part 1 (#16028), so I've recorded that as a blocker. It is not a duplicate of #16028. Nothing else is missing.
The work-resolve-issue lane stopped on this issue (host-only): the run reports the work needs a host run the lane cannot make.
It added host-run and set Work status to confirmed: an agent run on a host builds it, and it waits on no decision from a person.
In the run's words:
host-only: this issue's only done-when item is tracker state, not a file change. It needs one follow-on issue filed per plugin whose skill text changes, under container #15703, with the count stated on this issue. This lane only produces a working-tree patch and has no tracker write access (no gh and no work-item-tracker add), and it cannot read part 1's pull request body, which holds the keep-or-change list the follow-ons come from. A host run with tracker write access needs to read that list, file the items with /work-items:track add, and comment the count here.
What the run says it did not do:
Filing the per-plugin follow-on items under #15703 and stating their count on #16029: this lane has no tracker write access and cannot read part 1's PR body list
Follow-on items for the skill-text changes: 9, one per plugin, under #15703.
The list in #16602 routes 12 skills to "call /corrections:codify, presence-gated, with the strongest-first fallback". They sit in 9 plugins: instruction-placement 2, harness-config 3, and one each in harness-memory, docs-hygiene, testing, docs-naming, performance, planning and knowledge (2 + 3 + 7 = 12).
Each item carries its run budget. #15729's "three runs per side, and no more than 240 runs per plugin" is replaced by the owner decision of 2026-11-16: one targeted before-and-after case per changed skill on sonnet at 3 runs a side, with opus only to confirm a sonnet loss.
Report from #17200's host step for #17158 (pull request #17326, merged), on the recheck rule "If the recheck finds none for a skill, that skill is left unchanged and the finding is reported on #16029."
Two of the three changed harness-config skills had a sentence that picked a mechanism from a list (audit-instructions I5 and I15, draft-auto-mode-rules Phase 2 item 1); both are quoted in the pull request body.
audit-prompting-postures is the unclear one. It had no list. Its verdict rule read "MISSING (finding, with proposed addition as a fenced diff)", which gives every MISSING posture one fixed mechanism: a text addition to the audited file. The pull request changed the skill anyway (a P7 proposal line and an exception in the verdict rule).
Reading taken: that sentence is the mechanism-picking sentence, so the change stands. Evidence: in the host eval the grader hands-to-codify read 0 of 3 on the old text and 3 of 3 on the new text for this skill (record).
Done-when met: the nine follow-on items exist under #15703 and their count (9) is stated in the comment of 2026-11-18 above. Eight are closed by merged pull requests; #17160 is open with pull request #17548. The remaining host steps for the follow-ons (an eval case and a pull request body quote each) are tracked on #17200, and the open reading on audit-prompting-postures stays reported above for the owner.
Activity
github-actions commentedon Nov 13, 2026
Size: Small
Complexity: Low
Uncertainty: Knowable
Flags: none
Blocked by: #16028
I read this as part 2 of the split of #15729. Its one done-when item is that the follow-on items for the skill-text changes exist under the container, and that their count is stated on this item. It looks like it depends on the list of enforcement-placing skills from part 1 (#16028), so I've recorded that as a blocker. It is not a duplicate of #16028. Nothing else is missing.
Automated triage. A maintainer reviews it.
melodic-automation-lanes commentedon Nov 13, 2026
Not admitted to the issue lane yet: blocker #16028 is not closed by a listed human or a merge.
melodic-automation-lanes commentedon Nov 17, 2026
{"v":1,"actor_id":337595936,"title_sha256":"78ae2f3ea1230968ef769ec52dc31e7ac6537cc9abe9eae1d53179b8921399f4","body_sha256":"8efe040888da7a9d8c2acf05607cb0690776e551ccc53f3b367bf92ed3727fab"}
github-actions commentedon Nov 17, 2026
melodic-automation-lanes commentedon Nov 17, 2026
The work-resolve-issue lane stopped on this issue (
host-only): the run reports the work needs a host run the lane cannot make.It added
host-runand set Work status to confirmed: an agent run on a host builds it, and it waits on no decision from a person.In the run's words:
What the run says it did not do:
Run: https://github.com/melodic-software/claude-code-plugins/actions/runs/43997675480
To hand it back to the lane, remove
host-runand set Work status to agent-ready again.kyle-sexton commentedon Nov 18, 2026
Follow-on items for the skill-text changes: 9, one per plugin, under #15703.
The list in #16602 routes 12 skills to "call
/corrections:codify, presence-gated, with the strongest-first fallback". They sit in 9 plugins: instruction-placement 2, harness-config 3, and one each in harness-memory, docs-hygiene, testing, docs-naming, performance, planning and knowledge (2 + 3 + 7 = 12).Each item carries its run budget. #15729's "three runs per side, and no more than 240 runs per plugin" is replaced by the owner decision of 2026-11-16: one targeted before-and-after case per changed skill on sonnet at 3 runs a side, with opus only to confirm a sonnet loss.
kyle-sexton commentedon Nov 19, 2026
Report from #17200's host step for #17158 (pull request #17326, merged), on the recheck rule "If the recheck finds none for a skill, that skill is left unchanged and the finding is reported on #16029."
Two of the three changed harness-config skills had a sentence that picked a mechanism from a list (audit-instructions I5 and I15, draft-auto-mode-rules Phase 2 item 1); both are quoted in the pull request body.
audit-prompting-postures is the unclear one. It had no list. Its verdict rule read "
MISSING(finding, with proposed addition as a fenced diff)", which gives every MISSING posture one fixed mechanism: a text addition to the audited file. The pull request changed the skill anyway (a P7 proposal line and an exception in the verdict rule).hands-to-codifyread 0 of 3 on the old text and 3 of 3 on the new text for this skill (record).The #17158 entry on #17200 is ticked on the first reading.
Posted by an agent session.
kyle-sexton commentedon Nov 20, 2026
Done-when met: the nine follow-on items exist under #15703 and their count (9) is stated in the comment of 2026-11-18 above. Eight are closed by merged pull requests; #17160 is open with pull request #17548. The remaining host steps for the follow-ons (an eval case and a pull request body quote each) are tracked on #17200, and the open reading on audit-prompting-postures stays reported above for the owner.