03b7125293
* test(#3909): failing-first suite for the fabricated probe fallbacks Binds the four fabrication sites found by executing the surfaces (ADR-3889 failure class (c)), each with a positive control so an over-firing fix goes red: - the blocking api-coverage.verify-pre gate certifying "no external-API integration" from a zero-byte phase scope - the assumption-delta query route scanning an unresolvable phase section as the empty string and reporting it as an examined negative - both capability fragments' probe fallbacks, which append a fabricated verdict rather than replacing, and fire on the legitimate exit-1 negative Verification runs on the remote runner. Refs #3909 * enhance(#3909): a probe that could not run no longer asserts a verdict ADR-3889 Phase 5. Four sites turned a failed or unexamined probe into a confident negative; each now reports what it could not establish. - check api-coverage.verify-pre: a phase with no plan body and no roadmap section ran detection over zero bytes and PASSED the blocking seal gate, certifying "no external-API integration" from input it never read. It now holds with scope_unavailable. The discriminator is bytes examined, never signals found, so a phase whose plans are real and simply carry no API vocabulary passes exactly as before. - query assumption-delta scan: an unresolvable phase section was scanned as the empty string and reported as an examined negative. It now returns {skipped, reason: phase_unresolved}, still at exit 0 — an ADR-2980 degraded result in the payload, leaving the gsd-tools exit projection to P8. - both capability fragments: `|| echo '{"detected":false}'` appended rather than replaced, and fired on the legitimate exit-1 negative, so a correct answer and an honest skip both arrived as two concatenated objects. They now keep the probe's own payload and manufacture only an explicit probe_unavailable skip when the probe produced nothing at all. Every registered outcome is more restrictive on a blocking gate, so this can turn a false green red and never a red green. Docs: FEATURES 156, CONFIGURATION (both keys), references/api-coverage.md seal-time outcome table, and a new how-to for the reason-code vocabulary. Verification runs on the remote runner. Closes #3909 * test(#3909): correct the stale unknown-phase assertion `unknown phase → detected:false, no throw (graceful)` scanned phase 999 against a two-phase roadmap and asserted `detected === false`. That pinned the fabrication as intended behavior: the phase does not exist, so the detector was handed the empty string and its "no core assumption changed" answer described nothing that was ever read. It now asserts the skipped-with-reason shape. The graceful-degradation contract the test was actually protecting — the query succeeds and does not throw on an unknown phase — is unchanged. Found by code review, not by the author. Refs #3909 * docs(#3909): author the FEATURES entry in its generator source `docs/FEATURES.md` is generated by `scripts/gen-features.cjs` from the per-feature fragments in `docs/features/`. The API-coverage entry was edited in the generated file, so the next regeneration silently dropped it. The text now lives in `docs/features/api-coverage-gate.md` and `docs/FEATURES.md` is regenerated from it, leaving the shipped file byte-identical and its content actually derivable. Caught by `lint:generated-sync`. Refs #3909 * test(#3909): bind the skip to "not found", and pin the discriminator The first verification run went red on one case, and the case was wrong rather than the code. `getRoadmapPhaseWithFallback` returns `null` for an unknown phase and for a missing ROADMAP.md, but for a section whose body is whitespace-only it returns the heading line alone — which is not empty. So a body-less section WAS found, and reporting `detected:false` over its heading is a real negative, not a fabrication. The test had assumed the resolver yielded `''` there. Correcting the test rather than the resolver keeps `skipped` bound to the distinction the issue asks for — found versus not found — and avoids diverging `assumption-delta scan` from `roadmap.get-phase`, which the fragment documents as sharing one resolver. Also adds the seeded property the test matrix had promised: for any plan body, the scope read back is whitespace-only exactly when the body was. That pins the gate's discriminator to bytes examined, so it cannot quietly become "no signals found", across unicode whitespace and CRLF. `docs/INVENTORY.md` picks up the reference doc's new seal-time outcome table — surfaced by the co-change gate, not by a lint failure. Refs #3909 * chore(#3909): backfill the changeset PR number Refs #3909 --------- Co-authored-by: sim <sim@local>
57 lines
5.3 KiB
Markdown
57 lines
5.3 KiB
Markdown
# Assumption-Delta Architecture Checkpoint
|
|
|
|
> Advisory, non-blocking. Fires **only** when the phase scope shows a singular→plural / required→optional / derived→chosen transition. When it fires, it surfaces ONE identity-model question before the plan is finalized. Most phases will not fire it — that is the point.
|
|
|
|
## Why this exists
|
|
|
|
Most quietly-imported architectural debt does not come from a missing upfront design phase. It comes at the *seam*: a later phase introduces a second case (a second platform, auth method, tenant, region, source of truth) and nobody re-asks whether the original abstraction still names the right thing. The phase that adds the second case is exactly the 20-minute conversation that prevents an afternoon of later cleanup.
|
|
|
|
## Run the detector
|
|
|
|
The detector is a deterministic scan over the phase scope text. It strips fenced code blocks first, so a trigger word that appears only inside a code snippet does not fire. It returns a typed result: `{ detected, signals[], terms }`. Resolve it through the `assumption-delta scan` query (same phase-section resolver as `roadmap.get-phase`):
|
|
|
|
```bash
|
|
ASSUMPTION_DELTA_JSON=$(gsd_run query assumption-delta scan "${PHASE}" --json 2>/dev/null) || true
|
|
[ -n "$ASSUMPTION_DELTA_JSON" ] || ASSUMPTION_DELTA_JSON='{"skipped":true,"reason":"probe_unavailable"}'
|
|
```
|
|
|
|
> If the phase section cannot be resolved (no `ROADMAP.md` / unknown phase, or a section with no body), the query emits `{ "skipped": true, "reason": "phase_unresolved" }` — **not** `detected:false`. A probe that never had input does not get to assert that this phase changes no core assumption. The checkpoint does not fire either way; the difference is that a skip is now distinguishable from a real negative. Do not block on it.
|
|
>
|
|
> Optional tuning — pass `--terms <comma-list>` to replace the curated pluralization cues for this project (the `optional`/`chosen` cues keep their defaults): `gsd_run query assumption-delta scan "${PHASE}" --json --terms second,alternative,fallback`.
|
|
|
|
## Decision branch
|
|
|
|
Read `ASSUMPTION_DELTA_JSON`. Act on `detected` only — do **not** pattern-match the human prose.
|
|
|
|
**If `skipped` is `true`:** the detector never examined a phase section — it could not resolve one (`phase_unresolved`) or could not run at all (`probe_unavailable`). Skip the checkpoint for this run rather than asserting a verdict about input that was never examined; do not raise it with the user. **Check for `skipped` before reading `detected`** — a skipped payload carries no `detected` key, and treating its absence as `false` re-creates the fabrication this branch exists to prevent.
|
|
|
|
**If `detected` is `false`:** this phase does not change a core assumption. Skip the checkpoint entirely and continue planning. Do not raise it with the user.
|
|
|
|
**If `detected` is `true`:** a core assumption may have lost its monopoly. The `signals[]` array tells you which family fired:
|
|
|
|
| `kind` | What changed | The question to answer |
|
|
|---|---|---|
|
|
| `pluralization` | A second X was introduced where there was one (second platform / auth method / tenant / region / source of truth) | Does the current primary key / identity model still name the right noun? |
|
|
| `optional` | A required / `only` field became optional | Is the field still the right anchor, or has the anchor moved? |
|
|
| `chosen` | A derived value became chosen, or a constant became a parameter | Has a configuration decision become a modeling decision? |
|
|
|
|
Before finalizing the plan, answer this for the user and record the decision explicitly:
|
|
|
|
> **Promote vs. add-alongside.** The usual correct move when a generalization occurs is to **promote** the new general representation to the primary and **demote** the old specific one to a detail of one variant — *not* to add the new one alongside the still-required old one. Adding alongside silently contradicts the generalized intent (a later variant that does not fit the old primary can be stored but never confirmed as a default).
|
|
|
|
Record the outcome in the PLAN.md front matter / a `<assumption_delta_decision>` block:
|
|
|
|
- The **noun** that is now primary (the generalized identity).
|
|
- The **decision**: `promote` | `add-alongside` | `no-change`, with a one-line rationale.
|
|
- If `add-alongside`: call it out as accepted debt and note what would force a later promote.
|
|
|
|
## Optional companion: an invariant test
|
|
|
|
When `detected` is `true`, suggest (do not require) a contract/invariant test that encodes the now-generalized intent — e.g. *"every confirmed default round-trips through the primary use-path, for every supported variant."* That test goes red the instant a future phase reintroduces the singular assumption, so the regression cannot land silently. If the user accepts, add the test as a task in the plan.
|
|
|
|
## Tuning the vocabulary (optional)
|
|
|
|
The trigger vocabulary is a curated, additive-only set in `gsd-core/bin/lib/assumption-delta.cjs` (`DEFAULT_ASSUMPTION_DELTA_TERMS`). Bare "or" is intentionally excluded — it is too common in prose and would make the gate fire constantly. To widen or narrow the cues for a project, override at the call site with `--terms <comma-list>` (replaces the pluralization cues; `optional`/`chosen` keep defaults). The whole checkpoint is toggleable via `workflow.assumption_delta` in `.planning/config.json`.
|
|
|
|
This checkpoint is advisory: it informs and records; it never blocks the phase.
|