Learning #13 — a handoff's predictions decay

[rmsharp]

What this is

One row appended to the starter-kit/SESSION_RUNNER.md Learnings table (was 1–12; appended, never renumbered). No other change to that file, and no other file touched except the ledger entry.

A handoff carries predictions as well as facts — and a prediction about state the handoff's own session later changed is the least reliable kind.

This is the Phase 3C output of the session that shipped v3.6 (PR #62). It is a docs-only addition: no principle, phase, gate, workstream or failure mode changes, and the FM count stays 27.

Why the existing rows don't cover it

Existing row	What it covers
Learning #6, FM #11	claims written from memory
Learnings #7, #10, #12	cross-references that go stale in the corpus

Neither covers the forward-looking claim — a state description whose expiry is caused by the writing session's own subsequent commits. That is a distinct failure surface, and it bit this repo directly.

The motivating case is this repo's own v3.6 ship

The predecessor receipt told the ship session to "expect one CHANGELOG.md union conflict". git merge upstream/main produced seven.

The cause is worth stating precisely, because it generalizes: Layer 8 and the pre-PR review fixes were committed to the PR branch after the point the fork had been ported from it. So five files the prediction never named had diverged — both dashboard twins, the unit suite, CLAUDE.md and README.md. The prediction was true when written and false when used, and the session that wrote it is the session that invalidated it.

Why this is worth a table row rather than a gotcha: a wrong prediction is worse than no prediction. It licenses the successor to resolve quickly rather than carefully — and a merge resolved on the belief that "only the ledger differs" discards code silently, with no failing check to catch it. The countermeasure is mechanical and cheap: diff the branch against the commit the other side was ported from, not against its tip.

The row also asks the writer of a handoff to name the state a prediction depends on, so the reader can tell at a glance whether it still holds.

Cross-reference sweep (Learnings #7/#10, applied to this change)

Changing the size of a numbered set is exactly the trigger those rows describe, so the sweep was run before commit:

• No live count-claim about the Learnings table exists anywhere in the corpus. The table was 1-11 / table is 1-12 tokens are confined to dated CHANGELOG.md entries and HANDOFFS.md receipts, which this repo leaves verbatim by design.
• The table caption states no size.
• Nothing else required updating.

Verification

• bin/tests.sh — 84 passed, 0 failed
• bin/check-links — OK, 82 relative links across 21 distributed markdown files
• tools/test_methodology_dashboard.py — 197/197 OK
• Learnings table parses as rows 1–13, every row 4-column

Notes

starter-kit/SESSION_RUNNER.md is bin/_manifest.py-distributed, so adopters receive this via bin/sync.

No new HANDOFFS.md receipt is opened here. This is Phase 3C of fork session S17, whose durable receipt already exists on the fork; opening a second receipt for one session would double-count it. The CHANGELOG.md entry records the action and points at that receipt.

Whether this warrants a version event is yours to call — it is a single distributed-doc row, and v3.6 was tagged the same day.

🤖 Generated with Claude Code