Skip to content

cso: don't block the remediation roadmap on a human - #2862

Open
EmilianoU26 wants to merge 1 commit into
garrytan:mainfrom
EmilianoU26:pr1-cso-roadmap-unattended
Open

cso: don't block the remediation roadmap on a human#2862
EmilianoU26 wants to merge 1 commit into
garrytan:mainfrom
EmilianoU26:pr1-cso-roadmap-unattended

Conversation

@EmilianoU26

Copy link
Copy Markdown

An unattended /cso audit stops at Finding 1 and delivers nothing. The roadmap
asks one AskUserQuestion per finding, so a spawned session — where nobody is
reading this turn — has no way past the first one. The findings are already
written by then; the roadmap is the last step, not a gate.

When no human will answer (SESSION_KIND: spawned, or an auto-decide result),
take the RECOMMENDATION as the disposition and mark it AUTO-CHOSEN (unconfirmed):, then close the report with those lines under Decisions still
owed a human
so the ones never confirmed stay visible.

Two rules keep that from becoming a licence:

  • The spawned-session exception still applies to irreversible options. If the
    recommendation is C) Accept risk, record D) Defer instead — a risk nobody
    accepted is deferred, not accepted.
  • Read-only still holds. A) Fix now NAMES the change; it does not make it.

Interactive and Conductor sessions are unchanged: ask, or render the prose brief.

Edited in cso/SKILL.md.tmpl; cso/SKILL.md is the regenerated output.

⚠️ This does not fit under the parity ratchet, and I could not make it fit
without your call. /cso has roughly 150 bytes of skeleton headroom; this adds
~1100, so the suite reports skeleton 62740 > maxSkeletonBytes 61800.

I tried the architecturally correct move first — carving it into a new
cso/sections/roadmap-dispositions.md so it loads on demand — and it still
fails: registering ANY new section costs ~380 bytes of skeleton for its manifest
index entry and its pointer, which is more than the headroom. Compressing the
block to ~150 bytes leaves the claim without the procedure that makes it
actionable.

So the trade is yours: bump the ratchet, or name something in the /cso
skeleton that should move to a section first. I did not raise the cap or trim
someone else's paragraph to make room.

🤖 Generated with Claude Code

An unattended `/cso` audit stops at Finding 1 and delivers nothing. The roadmap
asks one AskUserQuestion per finding, so a spawned session — where nobody is
reading this turn — has no way past the first one. The findings are already
written by then; the roadmap is the last step, not a gate.

When no human will answer (`SESSION_KIND: spawned`, or an auto-decide result),
take the RECOMMENDATION as the disposition and mark it `AUTO-CHOSEN
(unconfirmed):`, then close the report with those lines under **Decisions still
owed a human** so the ones never confirmed stay visible.

Two rules keep that from becoming a licence:

- The spawned-session exception still applies to irreversible options. If the
  recommendation is C) Accept risk, record D) Defer instead — a risk nobody
  accepted is deferred, not accepted.
- Read-only still holds. A) Fix now NAMES the change; it does not make it.

Interactive and Conductor sessions are unchanged: ask, or render the prose brief.

Edited in `cso/SKILL.md.tmpl`; `cso/SKILL.md` is the regenerated output.

⚠️ This does not fit under the parity ratchet, and I could not make it fit
without your call. `/cso` has roughly 150 bytes of skeleton headroom; this adds
~1100, so the suite reports `skeleton 62740 > maxSkeletonBytes 61800`.

I tried the architecturally correct move first — carving it into a new
`cso/sections/roadmap-dispositions.md` so it loads on demand — and it still
fails: registering ANY new section costs ~380 bytes of skeleton for its manifest
index entry and its pointer, which is more than the headroom. Compressing the
block to ~150 bytes leaves the claim without the procedure that makes it
actionable.

So the trade is yours: bump the ratchet, or name something in the `/cso`
skeleton that should move to a section first. I did not raise the cap or trim
someone else's paragraph to make room.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@trunk-io

trunk-io Bot commented Sep 13, 2026

Copy link
Copy Markdown

Merging to main in this repository is managed by Trunk.

  • To merge this pull request, check the box to the left or comment /trunk merge below.

After your PR is submitted to the merge queue, this comment will be automatically updated with its status. If the PR fails, failure details will also be posted here

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.

1 participant