docs(feature-research/kilo): mirror hooks phrasing for subagents row

[Chen17-sq]

Summary

Addresses item 3 from #1616.

The subagents row in .rulesync/skills/rulesync-feature-research/references/kilo.md previously read:

subagents | https://kilo.ai/docs/customize/custom-subagents | Custom modes/agents; no README-supported Rulesync Kilo subagents target

That phrasing reads as if Rulesync has no support for Kilo subagents at all. In fact, KiloSubagent is registered in src/features/subagents/subagents-processor.ts (import { KiloSubagent } … at line 22, registered at line 184) — what's actually true is that the README's "Supported Tools and Features" table does not yet claim the target for Kilo.

The hooks row on the same file already encodes that exact nuance:

hooks | … | Adapter exists in source, but no README-supported Rulesync Kilo hooks target

This PR reuses that hedged phrasing for the subagents row, so the two parallel cases now read consistently. The Official docs URL is unchanged.

Test plan

• Single-line edit to one map file; no schema or generator changes.
• pnpm cicheck (fmt + oxlint + eslint + tsgo + 5550 vitest + sync-skill-docs check + cspell + secretlint) passes locally on Node 22.
• Pre-commit pnpm dev generate also runs cleanly, so the generated mirrors under .claude/, .opencode/, .github/, .takt/ regenerate without complaint.

Refs #1616 (item 3 of 5 — the other follow-ups in that issue are out of scope here).

[github-actions]

Thank you for your contribution! Unfortunately, you currently have 4 open PRs (including this one), which exceeds the limit of 2 for external contributors.

Please wait for an existing PR to be reviewed/merged, or close one before opening a new one. See CONTRIBUTING.md for details.

[Chen17-sq]

Closing this proactively — the github-actions bot flagged that I'm well over the 2-PR limit for external contributors documented in CONTRIBUTING.md. I opened these nine PRs in one batch without spotting that rule first, which was my mistake — sorry for the noise.

The patch in this PR is ready as-is (pnpm cicheck green, lint-staged green, pre-commit pnpm dev generate green) and stays small enough for a quick read; reopening is one click.

Keeping #1643 (refactor(utils): consolidate isPlainObject, addresses #1638) and #1644 (refactor(copilot-rule): reuse toPosixPath, addresses #1603) as my current two. I'll reopen this one — addressing #1616 item 3 — once a slot frees up.

[Chen17-sq]

Reopening — #1644 (toPosixPath copilot-rule refactor) merged 5/20, so I'm back at 1/2 PR limit. This is the smallest of the six remaining closed follow-ups: 1-line phrasing fix to mirror the existing hooks-row sentinel in the same file. Addresses #1616 item 3. Branch unchanged at fix/1616-kilo-subagents-wording @ 0f91bb1, mergeable, cicheck green when last run.

[dyoshikawa-claw]

The change itself is correct: removing the "Custom modes/agents; " prefix from the subagents description so it reads identically to the hooks row. That's a clean consistency win.

The only issue is the merge conflict. Commit 7dfac8e on main already updated this same line to include the "adapter exists" phrasing (but left the "Custom modes/agents; " prefix in place — that hybrid wording is what your PR addresses). A rebase onto current main should resolve trivially: the PR branch just needs to drop the prefix from main's current version of the line.

[dyoshikawa-claw]

The wording change matches the pattern already established by the hooks row — "Adapter exists in source, but no README-supported Rulesync Kilo ... target". No concerns with the content itself.

[dyoshikawa]

Thanks for this. The underlying observation is correct and the factual claim checks out — KiloSubagent is wired up in src/features/subagents/subagents-processor.ts and kilo-subagent.ts exists, so "adapter exists in source" is accurate. The issue is that main already moved this line: it now reads "Custom modes/agents; adapter exists in source, but no README-supported Rulesync Kilo subagents target", which already folds in the nuance you're after. The PR is also CONFLICTING against main now. On top of that, this PR's wording drops the "Custom modes/agents;" upstream-surface prefix (see inline), which main keeps for a reason — unlike the hooks row, the subagents row does have an upstream surface. So mirroring hooks exactly loses information here. I'd suggest closing this; the intent is already on main. If you want to keep the branch, rebase and preserve the "Custom modes/agents;" prefix.

[dyoshikawa]

This drops the leading "Custom modes/agents;" prefix that main keeps. The hooks row omits an upstream-surface phrase because it has no dedicated upstream surface ("No dedicated upstream hooks surface in map"), but the subagents row does point at an upstream surface (custom-subagents) realized as custom modes/agents. So aligning the two phrasings verbatim removes meaningful upstream info. Better to keep the prefix.