Repository navigation
Description
Problem
The melodic-software skills overflow the shared skill-listing budget on 200k-window models, and
the fix is at the source: shorter descriptions. Two follow-ups have landed. PR #4376 (claude-ops
0.59.1) fixed audit-skill-visibility --installed counting never-enabled plugins as enabled.
PR #4555 (discipline 0.14.9, #4552) trimmed 13 discipline descriptions to 500 characters or fewer.
The work below is what remains, split into three child issues:
- Trim descriptions in the other heavy plugins (child: description trims).
- Decide whether any skill should take
disable-model-invocation: trueto save listing cost,
which conflicts with a standing convention (child: needs-decision). - Align the remaining surfaces with the observed enablement rule: an installed plugin that no
enabledPluginsscope names does not load. The surfaces arelevers.json:257, the claude-ops
fleet-state docs, and the non-BooleanenabledPluginshandling (child: enablement rule).
Evidence
Measured numbers, all in source description characters (description + when_to_use, skills
without disable-model-invocation: true):
bash plugins/skill-quality/scripts/check-listing-budget.sh plugins/*/skillsata6ba321e1
(2026-09-27): 208 listing-eligible skills across 80 roots, 152,297 chars. That is about 19x the
8,000-char default the script assumes, and the script calls it an estimate. A separate parse
at origin/main946caf11gives 152,182.- The operator machine measured with
audit-skill-visibility --installedon Claude Code 2.1.270
(item, Part C). The 63 explicitly enabled plugins had 166 model-visible skills totaling 108,992
chars. Budget atskillListingBudgetFraction: 0.05on a 1M window, at 4 bytes per token, is
200,000 chars, so the listing fits. On a 200k window it is 40,000 chars and overflows by 68,992.
At the default 0.01 on a 1M window it is also 40,000. A live 1M session at 0.05 showed every
claude-ops description in full, so the fit was observed, not only computed. - Fitting 108,992 chars into a 200k window would need a fraction of about 0.14 to 0.18, so
raising the fraction does not help there. Source trims do. That case covers sonnet and haiku
subagents. skillListingMaxDescCharsdefault is 1,536, so no description is capped; every character
counts.- The settings reference now documents the
skillListingBudgetFractiondefault as0.01("which
reserves 1% of the context window"), re-fetched 2026-09-27. The item said it had no documented
default; that is superseded. - The item's Plugins > Stats tab figures (about 15,260, about 5,500, about 1,960) have no
documented unit. Do not use them or convert them to tokens.
The per-plugin table is in the trims child.
Proposed approach
Work the three children independently. The operator decided on 2026-09-25 (inbox loop h17):
- Trim the listing per plugin, with no renames and no skill merges.
- Never set
disable-model-invocationon a skill that the PR ready flow,extract-ssot, or
audit-passinvokes through the Skill tool.
The item also proposed three things that this decision rules out. Consolidating the four
docs-hygiene:audit-* skills into one moded audit is a merge, so it is declined. Folding the
-deep discipline variants into their base skills is also a merge, so it is declined. Disabling
unused vendor plugins to save listing cost is dropped because those plugins never loaded.
Acceptance criteria
- All three child issues closed or explicitly declined.
-
check-listing-budget.sh plugins/*/skillsaggregate recorded before and after in each trims
PR.
Constraints and gotchas
- "Never invoked" does not prove a skill is useless. This is a deliberate pass, not bulk deletion.
- Leave these to other lanes, not this repo:
- The dotfiles reconciliation (the item's Part B) duplicates local item 20260912-222500.
- The
autoMode.classifyAllShellprobe needs an SW2030 session.
Context
Source: local handoff item 20260923-182500-skill-listing-context-cost-remainder-after-4376.md,
retired into this issue and its children.
Related:
- PR fix(claude-ops): count never-enabled plugins as not-enabled in audit-skill-visibility #4376, audit-skill-visibility: --installed counts never-enabled plugins as loaded, inflating the fleet ~3x and every starvation verdict with it #4150, fix(claude-ops): audit-skill-visibility counts never-enabled plugins as enabled #4375 (never-enabled miscount)
- PR perf(discipline): trim skill descriptions to 500 characters for the shared listing #4555, discipline: skill descriptions are the largest melodic share of the shared skill listing #4552 (discipline trims)
- perf(skills): skill listing is 42% over its character budget; ~50 skills lose their descriptions and become undiscoverable #3505, skill-quality: measure and improve description-driven auto-invocation (eval-based) #3526
docs/conventions/invocation-mode/README.md
Activity
perf(docs-hygiene): trim skill descriptions under 500 chars (part of #…
perf(work-items): trim skill descriptions to 500-char listing target (#…
perf(planning): trim skill descriptions; keep extractor triggers (par…
perf(claude-config): trim skill descriptions under 500 chars (part of #…
perf(planning): trim skill descriptions; keep extractor triggers (par…
perf(claude-ops): trim skill descriptions under 500 chars (part of #4657
perf(knowledge): trim skill descriptions (#4661) (#4955)
perf(planning): trim skill descriptions under 500 chars (#4657) (#4931)
perf(docs-hygiene): trim skill descriptions under 500 chars (#4657) (#…
docs(out-of-scope): record skill-listing 500-char exceptions (#4661, #…
perf(overengineering): trim skill descriptions (#4661) (#4958)
perf(code-tidying): trim skill descriptions under 500 chars (#4661) (#…
6 remaining items
fix(claude-ops): name /skill-quality:check in the native-overlap Appl…
docs(out-of-scope): delete the skill-listing 500-character exceptions…
docs(conventions): drop the 500-character ledger pointer; say no conf…
docs(work-items): point triage exits at the lane-barred branch and co…
kyle-sexton commentedon Sep 29, 2026
From the 2026-09-29 audit of the Cursor run. Reopened.
#4774's body said "Does not close #4657", yet the epic auto-closed at 64cdc57 (github-actions warned 13 minutes before merge). AC1 (all three children closed or declined) did not hold then. State today (
gh issue view, 2026-09-29): #4659 closed 2026-09-28, #4660 closed 2026-09-29 (by #5202), #4661 still open. The audit report's "link #4660 as its tracker" is stale.So AC1 does not yet hold. Reopened until #4661 is closed or declined. Children: #4659, #4660, #4661. No code change here.
fix(claude-ops): correct doctor claims, changelog, lanes and prerequi…
docs(conventions): revert agent rulings on owner-reserved issues and …
fix(planning): audit remediation (interview scope, outcome gate, plan…
fix(work-items): resolve audit findings for work-loop, adapter releas…
perf(skills): trim architecture, improvement, claude-config, discipli…
docs: remove agent-encoded decisions and correct decision records (#5288
perf(skills): trim the remaining listed descriptions over 500 in #4661 …
kyle-sexton commentedon Oct 1, 2026
Owner decision (2026-10-01): accepted the recommendation ("accept all recommendations" and "I want you to do all of this for me", the owner's reply to the list that named this item).
Criterion 2 is waived: seven merged trims PRs (#4776, #4930, #4931, #4954, #4955, #4958, #5258) never recorded the before/after listing-budget aggregate, and those PR bodies are history. Every child (#4659, #4660, #4661) is closed as completed.