[AUTOMATED] docs(re-needs): 10 needs were closed without an acceptance flip — say so - #423
Merged
Merged
Conversation
…e flip — say so Auditing which need fields are actually written, 10 of 51 records carry no `probe_id`. All ten are records I filed BY HAND -- the eight round-1 needs written from ADMITTED.md and two round-2 seeds. Every need the loop filed through cluster.py has both arms. All ten are `closed`. None was closed by the gate, because none has an acceptance probe to flip: a human decided they were done. That sits outside the design's central rule, and the contrast is sharp -- the captain enforced exactly that rule on `decompiling-3396-byte-main`, refusing six times to relax a 10 s bar rather than redefine `closed` as "a builder tried". The capabilities are real and were verified by hand at the time. The point is that the RECORD does not carry the evidence, so the closure is unauditable and these needs are invisible to regression detection. Each now says so under `## Acceptance` and names the durable artefact that does cover it (tests/cli probes, tests/stages cases, or the PR). DELIBERATELY NOT BACK-FILLED. A probe written now would assert today's behaviour, not the behaviour at closing time, and would sit in the record looking like gate evidence while being nothing of the kind. An honest gap beats a manufactured flip. The note lives under `## Acceptance` rather than a new `## Evidence` heading because SECTIONS is a fixed list and the round-trip drops anything outside it -- caught by the smoke's own front-matter round-trip, 10 failures, one per record I had just edited. Rebuilt from origin/main after a first attempt's `git add docs/re-needs/` swept up 17 round-3 records the concurrently-running loop had written but not yet committed. This commit touches exactly the ten files it describes. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
mahaloz
force-pushed
the
docs/link-handfiled-evidence
branch
from
September 5, 2026 22:33
8b3e8dc to
f38b3e3
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Auditing which need fields are actually written, 10 of 51 records carry no
probe_idand 8 noacceptance_id. All ten are records I filed BY HAND -- theeight round-1 needs written from ADMITTED.md and two round-2 seeds. Every need the
loop itself filed through cluster.py has both.
All ten are
closed. None was closed by the gate, because none has an acceptanceprobe to flip. A human decided they were done. The design's central rule is that
only the acceptance probe may close a need -- the captain enforced exactly this on
decompiling-3396-byte-main, refusing six times to relax a 10 s bar rather thanredefine
closedas "a builder tried" -- and these ten quietly sit outside it.The capabilities are real and I verified them by hand at the time; the point is that
the RECORD does not carry that evidence, so the closure is unauditable and the needs
are invisible to regression detection. Each now says so under
## Acceptance, andnames the durable artefact that does cover it:
no-strings-inventory tests/cli/strings-inventory-with-xrefs.json, ...
no-disassembly-command tests/cli/disassemble-named-function.json
whole-binary-json-untriageable tests/cli/functions-summary-is-triageable.json
phantom-unmapped-entry tests/cli/unmappedentry-no-phantom-entry.json + stage
main-void-prototype tests/stages/kuna-entrymainproto.xml
switch-on-constant-selector tests/stages/kuna-switchselector.xml
int80-opaque-syscall tests/stages/kuna-linuxsyscall.xml
call-args-discarded tests/stages/kuna-calleearity.xml, -varargstackargs.xml
no-cli-data-code-override #391
no-cli-rename-or-prototype-... #389
DELIBERATELY NOT BACK-FILLED. A probe written now would assert today's behaviour, not
the behaviour at closing time, and would sit in the record looking like gate evidence
while being nothing of the kind. An honest gap beats a manufactured flip.
The note lives under
## Acceptancerather than a new## Evidenceheading becauseSECTIONS is a fixed list and the round-trip drops anything outside it -- caught by the
smoke's own front-matter round-trip, 10 failures, one per record I had just edited.
tools/repipe/smoke.sh 127/127.
Co-Authored-By: Claude Opus 5 (1M context) noreply@anthropic.com
🤖 Generated with Claude Code
https://claude.ai/code/session_01YcFmfZndNjgfqLQVBZCdkY