FE-674: Plan V3.1 verification strategy + node-edit cards#119
Open
kostandinang wants to merge 2 commits into
Open
FE-674: Plan V3.1 verification strategy + node-edit cards#119kostandinang wants to merge 2 commits into
kostandinang wants to merge 2 commits into
Conversation
This was referenced May 8, 2026
Contributor
Author
This was referenced May 8, 2026
kostandinang
added a commit
that referenced
this pull request
May 8, 2026
memory/CARDS.md is the per-branch execution queue for the next three
buildable scope cards on this frontier (ka/fe-674-cascade-edit-and-agent):
1. Source-content snapshots on reconciliation_need (server)
- Add source_previous_content + source_current_content TEXT NULL
- Thread through OpenReconciliationNeedInput + edit-route hard path
- Expose on ReconciliationNeedRecord shared type
2. Source diff rendered inline on each Pending review row (client)
- Reuse <ContentDiff> from FE-665
- Render only when both snapshots are present and non-equal
- Legacy rows (snapshots null) keep today's bare layout
3. Edit-target affordance per Pending review row (client + reuse)
- Inline textarea pre-filled with target's current content
- Saves through existing PATCH /knowledge-items/:id then resolve endpoint
- Re-entrant cascade works without reload
All three are light scope cards; promotion checklist run on each, all
stay light (no new requirements, assumptions, decisions, invariants;
all stay inside the V3.0 cascade seam).
The V3.1 agent slices (4-6) are deliberately NOT pre-queued -- they
depend on Card 3's inline-edit shape and on the LLM oracle strategy
landing in PR #119 (ka/fe-674-plan-v3-1-verification). Re-run ln-scope
for slice 4 after Card 3 ships.
CARDS.md follows AGENTS.md's sanctioned exception: it is a derivative
queue scoped to one branch and one frontier item, not canonical
narrative -- so it lives on the feature branch, not the planning PR
(per /planning-pr decision rule).
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
51310d0 to
a492dcf
Compare
3 tasks
PR SummaryLow Risk Overview Extends Reviewed by Cursor Bugbot for commit 466e4cb. Bugbot is set up for automated code reviews on this repo. Configure here. |
🤖 Augment PR SummarySummary: Updates the V3.1 reconciliation verification plan and execution sequencing in the canonical planning docs. Changes:
🤖 Was this summary useful? React with 👍 or 👎 |
![augmentcode[bot]](https://avatars.githubusercontent.com/in/1027498?s=60&v=4){kind=small}
kostandinang
changed the base branch from
ka/fe-674-cascade-polish
to
graphite-base/119
kostandinang
force-pushed
the
graphite-base/119
branch
from
9464e9f to
f74f37c
Compare
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
a492dcf to
7fc63be
Compare
kostandinang
changed the base branch from
graphite-base/119
to
ka/fe-674-cascade-polish
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
9ed2e7b to
9e27baa
Compare
kostandinang
force-pushed
the
ka/fe-674-cascade-polish
branch
from
f74f37c to
37085de
Compare
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
9e27baa to
642f4f7
Compare
kostandinang
force-pushed
the
ka/fe-674-cascade-polish
branch
2 times, most recently
from
41359bd to
675db5b
Compare
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
642f4f7 to
387cbb5
Compare
kostandinang
force-pushed
the
ka/fe-674-cascade-polish
branch
from
675db5b to
bfeff81
Compare
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
387cbb5 to
95f6976
Compare
Updates memory/SPEC.md §Verification Design with the three-ring oracle
strategy for the side-chat V3.1 reconciliation classifier, and rewrites
the V3.1 entry in memory/PLAN.md §Next to fold in two adjacent
node-edit cards that ship before the agent itself.
memory/SPEC.md
- §Oracle Strategy by Loop Tier: 3 new rows for V3.1
- Inner: deterministic state-machine tests over agent_status with a
stubbed classifier (queue lifecycle + label vocabulary)
- Middle: golden-fixture corpus of (source change, target content,
relation kind) -> expected classification, regressable across
prompt revisions
- Outer: manual cascade walkthroughs on dense specs comparing
agent-grouped vs flat-list resolution -- the only ring that can
validate A88 ("does grouping help")
- §Design Notes: 2 new notes
- Three-ring rationale for V3.1 classifier verification, with the
recoverability argument (agent_proposal never auto-applied) that
lets the inner/middle rings stay shallow
- Cascade-edit cards (snapshots + inline target edit) are inner-loop
only; promote a route/query ownership oracle if re-entry surfaces
issues under manual walkthrough
- §Acknowledged Blind Spots: 4 new entries
- V3.1 classifier label correctness on substantive items (highest
stakes; recoverable but not testable without ground-truth corpus)
- V3.1 classifier multi-run determinism (single-shot first cut, no
temperature pinning or variance characterization)
- V3.1 cross-need correlation (per-need independence; agent cannot
see shared root concepts)
- Re-entrant cascade behavior under inline target edit (Card 3)
No new A##, D###, I###, or Requirement N rows -- existing IDs cited
but not modified, so the planning-pr hard rule on tracked-ID rows
does not fire. Verification Design is paragraph/table-row level.
memory/PLAN.md
- V3.1 Next entry: rewritten description folds in node-edit cards;
explicit Linear note (FE-674, separate stacked branch); explicit
"user has elected to proceed without V3.0 corpus signal" note.
- Adds **Verification approach** with the three-ring summary and
cross-reference to SPEC.md §Verification Design.
- Adds Build order: Cards 1-3 (snapshots, inline source diff,
Edit-target affordance) ship first; re-run ln-scope for the agent
backend slice after Card 3.
Per the /planning-pr skill (FE-703): paragraph-level rewrites of
PLAN.md and Verification Design changes go on a planning PR. The
planning-pr default base is main; per user direction this branch
stacks on ka/fe-674-cascade-polish so the V3.1 work stays under
FE-674 rather than spawning a sibling Linear issue.
memory/CARDS.md (the per-branch execution queue for Cards 1-3) is
deliberately NOT included here -- it lives on the downstream feature
branch ka/fe-674-cascade-edit-and-agent as a per-frontier artifact,
not canonical narrative.
Verification: pre-existing structured-list-view.test.tsx failures on
the parent branch baseline (cascade-polish) are unrelated to this
memory-only change; full test suite otherwise green.
Add the I114 invariant row for the V3.1 reconciliation-agent classifier lifecycle (the two existing 'planned I114' references in SPEC's verification table were dangling). Spell out in PLAN that agent_status is a separate V3.1 field from the durable status (open/resolved) shipped in V3.0.
kostandinang
force-pushed
the
ka/fe-674-cascade-polish
branch
from
bfeff81 to
d2a3611
Compare
kostandinang
force-pushed
the
ka/fe-674-plan-v3-1-verification
branch
from
95f6976 to
466e4cb
Compare
6 tasks
kostandinang
changed the title
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.
What
Planning-doc updates only (
memory/SPEC.md,memory/PLAN.md). Records the verification strategy and node-edit cards for V3.1. No code.Stacked on #118 per user direction (default for planning PRs is
main).Changes
Test plan