Skip to content

docs(rfc): record the Slack Code channel decisions the next pull requests build on - #38124

Merged
kyleseaman merged 1 commit into
mainfrom
docs/rfc-slack-code-shape-decisions
Nov 17, 2026
Merged

kyleseaman merged 1 commit into
mainfrom
docs/rfc-slack-code-shape-decisions

Conversation

@kyleseaman

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Record the owner's decisions on what a Slack code channel shows and runs, on main, before the code pull requests that build them.

The First Principles lane blocks a code pull request that changes product shape unless the decision is already in an RFC on the base branch. An RFC edit inside the code pull request does not count; in its words on #22370, "Changing the RFC inside the PR only proposes the decision; it doesn't record it." Several Slack Code pull requests change what a code channel shows or runs, and docs/request-for-change/rfc-slack-code-channels.md (status partial) does not describe those changes on main 09e5482863.

Why it matters

Without the decisions on main, each of these pull requests is blocked until the owner posts an override, and posts it again after every push. One docs-only amendment, merged first, records them once.

Not a goal

  • Any code. No src/ or test/ file changes, and nothing is generated.
  • New decisions. Each section records a decision the owner already made, dated where the decision has a date, and describes what the code pull request builds, no more.
  • The 2026-11-12 Stop amendment, which stays as it is.
  • A docs/decisions entry: an RFC amendment does not need one (docs(rfc): a code channel's Stop also stops its helper agents #35935 had none).

What changed (motivation → approach → change)

The amendment follows #35935: a dated "Amended 2026-11-16" item under the document's status line that names each decision, its date, its section and what main has today, then the text in the sections it amends. It takes effect when the owner merges this amendment.

What each part records, and the pull request it unblocks:

Part of the RFC Decision Unblocks
§3 "Helper agents", Backward compatibility Helper agents (slack.code_channel_dispatch) turn on with a first enable of Code channels; a stored value, including an explicit off, is kept; a missing or unreadable config reads off. Decided 2026-11-13. The coming helper-agents pull request (branch feat/code-channel-helpers-default-on)
§3 "Always use conductor mode", "Resolution order", "The auto folder", Backward compatibility A code channel runs the owner's default agent unless told otherwise. Always use conductor mode (off by default, offered wherever conductor mode is) makes every new code channel run agent.crew_mode_agent, over the channel default agent and the folder's agent. The auto folder's legacy kirocrew-conductor reads as unset. The session carries no conductor mode stamp, so its helper agents stay filed in the channel's folder. Decided 2026-11-13; given the conductor mode name on 2026-11-16. The coming default-agent and conductor mode switch pull request
§3 "Context bar", "Plan-canvas comments", the manifest table, open question 2 Two action items (Create PR, Stop), a Workers item, one PR slot with CI in its label; Plan-canvas comments reach the agent; the code_channel_action event is subscribed. Decided 2026-10-23. #22370 (its First Principles block names the action items, the Workers item with one PR slot, and CI in the label)
§3 "Tabs" Native PR tabs, tab deep links, a Workers tab with a per-helper Stop and Open in dashboard, the five-tab limit and which tab gives way. The owner's design built in #22928; decided 2026-11-16 to ship on main as its own pull request. The coming tabs pull request, on main after #22370
§3 "Work board": "No reactions", "Filing" Records what #21993 shipped: bound code channels show no reaction ladder and no queued-message hourglass, progress is the session status and the pinned work board, DMs, threads and ordinary channels are unchanged (owner's decision, 2026-11-13); the board files no helper agents, session creation files them, and a channel whose session turned conductor mode on keeps them out of folders (2026-10-27 decision). Already built; recorded so later pull requests that touch the board have it on main
§3 "Work board": "PR and CI per row", "Combined Code tab", "Wrap-up outcomes", "Idle !stop" The parts of the work board #21993 left for a second pull request. The coming board pull request (branch feat/slack-code-channel-board-diff)

The front matter's last-audited, audited-at and implementation-prs (adds #21993) are brought up to date, and the index row in docs/request-for-change/README.md records the amendment and what main has. Each "on main" statement was checked in the code at 09e5482863: features.AUTO_FOLDER_AGENT is kirocrew-conductor, slack_code_settings.py never saves code_channel_dispatch on a Code channels enable, slack/code_channels/ has no actions.py, views.py, comments.py or workspace/worker_prs.py, the manifest subscribes no code_channel_action, and an idle !stop gives only a count of running helper agents.

Backwards compatibility

Compatible: documentation only; no code, config or behaviour changes.

Removes nothing: the change adds text and rewrites lines that described the old shape: open question 2's "link items only", the auto folder's kirocrew-conductor default and its "· kirocrew-conductor" settings label, the agent row of the resolution order, the context bar's "(≤5 items: repo, branch, PR, CI)", the code_channels manifest row's events, "With all of them off" (now "With every switch off"), and a qualifier after the status line's "so no default changes". No section or reader goes away.

Tests

Proof: not provable -- docs or config only

On this head: ./scripts/docs-lint.sh, scripts/docs_lint.py --test, the brand-name, comment-history, feature-map, changelog-history and decisions-history source gates, and in one run every test file that names request-for-change (42), test_docs_lint_fact_checks.py, test_interactions_options.py and every test that scans all Markdown docs (15): 3,141 passed, 2 skipped.

Manual verification

N/A — documentation only. Each section was checked against the code hunk of the pull request it unblocks.

Related Issues

None. This records decisions; the code pull requests follow.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

Contribution License Agreement

Contributions are accepted under the project's Apache-2.0 license.

@kyleseaman
kyleseaman requested review from a team and KiroCrewCaptain as code owners November 16, 2026 21:56
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 16, 2026
@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision c980649d744c48c5f94f4c20f68ab42687aef3ff touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ⏭️ skipped

Revision c980649d744c48c5f94f4c20f68ab42687aef3ff ships no reviewable capability (docs, tests or generated files only), so there is nothing to inventory. Does not block merge.

@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of c980649d744c48c5f94f4c20f68ab42687aef3ff — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

This is a docs-only RFC amendment that records decisions already made. It adds no code, removes no reader and changes no surface that renders.

[DESIGN-REVIEWED] c980649

Description read

[DESCRIPTION-READ] 53814c24fe9ed3275259f3b6fe0992e32011247e741c124e5ba810da7f4ba8bd

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed c980649d744c48c5f94f4c20f68ab42687aef3ff — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] c980649

Verdict parsed from the review's SHA-scoped output markers for commit c980649d744c48c5f94f4c20f68ab42687aef3ff.

False positive or not applicable? A repository writer can comment:
/ai-review override fable c980649d744c48c5f94f4c20f68ab42687aef3ff false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 16, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of c980649d744c48c5f94f4c20f68ab42687aef3ff and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] c980649

False positive or not applicable? A repository writer can comment:
/ai-review override gpt c980649d744c48c5f94f4c20f68ab42687aef3ff false-positive: <reason, 20 to 500 characters>

…ests build on

Amends rfc-slack-code-channels.md with the owner's decisions on helper agents
turning on with a first enable of Code channels, the default agent and Always use
conductor mode, the context bar's action items and single PR slot, Plan-canvas
comments, the PR and Workers tabs and the tab limit, and the work board's per-row
PR and CI, combined Code tab, wrap-up outcomes and idle !stop reply. Records what
#21993 shipped: no reactions in bound code channels, and the board files no
helper agents. Updates the RFC index row.
@kyleseaman
kyleseaman force-pushed the docs/rfc-slack-code-shape-decisions branch from 7cfd419 to c980649 Compare November 16, 2026 23:10
@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 17, 2026
@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: A person should decide

Why: the decider estimates a 10% chance that this change introduces a defect, within the 12% limit for approving on its own, but it is held at tier 2: design-doc-gate: D.

This PR changes docs only. It amends the Slack Code channels RFC and its README index row to record decisions the owner has already made, so the code PRs that follow can rely on them. I checked its "on main" claims against the head code and they hold:

  • features.AUTO_FOLDER_AGENT is kirocrew-conductor.
  • slack/code_channels/ has no actions.py, views.py or comments.py.
  • The manifest subscribes only agent_session_stopped for code channels.
  • The cited 2026-10-27, 2026-11-14 and 2026-11-15 decision files exist.

The trust sweep found nothing that fails open:

  • Context-bar actions refuse clicks they can't attribute or whose fields name conflicting people.
  • The Workers-tab Stop uses the same owner-only authorization as the channel's Stop, and the gateway resolves its target from the channel's own board.
  • Plan-canvas comments go through the same authorization as a message the commenter typed in the channel.
  • Turning helper agents on with the first enable of Code channels keeps any stored value, including an explicit off, and an unreadable or missing config reads as off. The owner-only and approval-mode limits are unchanged.
  • The manifest stays byte-identical with every

Tier: 2, held for a person: design-doc-gate: D.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit c980649d74 · 254 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

This changes behaviour users already rely on, so it needs a design sign-off from a maintainer before it can merge. A maintainer will post the decision here.

To propose a design, follow the request-for-change process in CONTRIBUTING.md (docs/request-for-change/).

Review result on this head: no blocking findings. The diff touches only the RFC and the request-for-change README. We checked each "on main" claim at the audited commit 09e5482, which is an ancestor of main. features.AUTO_FOLDER_AGENT is still kirocrew-conductor. slack/code_channels/ has no actions.py, views.py or comments.py. No source file names code_channel_action. #21993 is on main. The linked decision entries exist and agree with the text. The amendment keeps "decided" separate from "built". Backward compatibility now names both default changes and the manifest drift gap for the new event.

Optional:

@kyleseaman
kyleseaman merged commit c30bfba into main Nov 17, 2026
101 checks passed
@kyleseaman
kyleseaman deleted the docs/rfc-slack-code-shape-decisions branch November 17, 2026 02:27
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 17, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants