Files
msd-core/.changeset
Tom Boucher 75ee7b0214 enhance(#4273): add phase.tdd-applicable single-owner predicate (#4277)
* enhance(#4273): add phase.tdd-applicable single-owner predicate

One query verb computes TDD-applicability for a plan (CLI flag, plan
type: tdd frontmatter, a task's tdd="true" attribute, or the
workflow.tdd_mode config default), mirroring phase.mvp-mode's
precedence-cascade shape. Foundation for epic #4272 Phase 2, which
wires both dispatch backends to consume it instead of restating the
predicate independently.

Also fixes workflow.tdd_mode, workflow.research, and
workflow.nyquist_validation, which never reached
cmdInitExecutePhase/cmdInitPlanPhase/cmdInitDebug/cmdInitNewMilestone
because loadConfig() never populates config.workflow — a dead
accessor found while wiring this verb's own config read, fixed inline
per the no-defer rule rather than left alongside it.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* docs(#4273): document phase.tdd-applicable's FEATURES.md entry

Add a docs/features/ fragment for the new phase.tdd-applicable query
verb and regenerate docs/FEATURES.md. docs/COMMANDS.md is left
untouched: it documents /gsd-* slash commands only, and the sibling
verb phase.tdd-applicable mirrors (phase.mvp-mode) has no formal CLI
reference entry anywhere in docs/ either -- only inline prose mentions
in docs/reference/workflow-fragments.md -- so there is no COMMANDS.md
precedent to extend.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#4273): use PHASE_NOT_FOUND reason code, remove try/finally from tests

Two orthogonal code reviews flagged a mistyped error reason and a CONTRIBUTING.md-banned try/finally pattern in the phase.tdd-applicable change; both are corrected here.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#4273): stop whitelisting capability-owned config keys centrally

workflow.tdd_mode, workflow.research, and workflow.nyquist_validation are
each already owned by their own first-party capability's federated config
schema (the tdd/research/nyquist capabilities declare them under their own
capability.json `config`), resolved via isCapabilityConfigKey. Adding them
to gsd-core/bin/shared/config-schema.manifest.json's central validKeys, as
the prior commit in this branch did (mirroring workflow.mvp_mode, which
genuinely is central-only), declares the same key in two places at once.
That collision breaks capability-loader.cts's loadRegistry composition:
gsd-test caught this as 84-85 unrelated failures across
capability-cli/capability-command-dispatch/capability-lifecycle test files,
every one showing "unknown capability: <id>" for a freshly-installed
third-party capability that should have resolved fine.

Verified directly (not asserted): reverting only this file, keeping the
config-loader.cts tdd_mode/research/nyquist_validation flattening and the
init.cts call-site fixes from the prior commit, and re-running the exact
capability install + capability set repro from
tests/capability-cli.test.cjs's "issue-2322" test locally reproduces the
failure with the whitelist entries present and clears it without them.
loadConfig() still surfaces all three flattened values correctly with no
central whitelist entry (confirmed directly against the compiled module) —
the whitelist additions were never required for the #4273 fix to work; they
were an incorrect over-application of the mvp_mode precedent to keys that
aren't central.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* fix(#4273): use getNested for tdd_mode (no legacy top-level fallback), allowlist new test file

Both fixes address defects found by a gsd-test bench run: tdd_mode routed through get() invented an undocumented top-level alias that silently outranked the canonical workflow.tdd_mode key, and the new phase-tdd-applicable test file was missing from the file-count allowlist.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

* chore(#4273): backfill changeset PR number

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

---------

Co-authored-by: sim <sim@local>
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-09-04 14:14:56 -04:00
..

Changeset Fragments

This directory holds per-PR CHANGELOG fragments. Every PR with user-facing changes drops one (or more) <random-name>.md files here describing its CHANGELOG entry. Fragments are consolidated into the top-level CHANGELOG.md at release time.

Why

Two PRs that both edit the ### Fixed block of CHANGELOG.md always conflict on merge — git can't pick a serialization order without human input. Two PRs that each add a fresh .changeset/<unique-name>.md never conflict because they don't share lines.

See #2975 for the full rationale.

Adding a fragment

node scripts/changeset/new.cjs \
  --type Fixed \
  --pr 1234 \
  --body "fix the thing — explain the user-visible change in one sentence"

This writes .changeset/<adjective>-<noun>-<noun>.md with frontmatter and a body. Three random words → concurrent PRs don't collide.

Format

---
type: Fixed
pr: 1234
---
**`/gsd-foo` no longer drops trailing slashes** — explain the user-visible change.

Allowed type: values follow Keep a Changelog: Added, Changed, Deprecated, Removed, Fixed, Security.

Opting out

PRs that legitimately have no user-facing impact can add the no-changelog label. CI honors it. When unsure, add the fragment.

At release time

Promotion is automatic. The release workflow's finalize job runs:

node scripts/changeset/cli.cjs render --version vX.Y.Z --date YYYY-MM-DD --allow-empty

This reads every fragment, groups bullets by type:, replaces ## [Unreleased] with a new ## [vX.Y.Z] - YYYY-MM-DD block, opens a fresh ## [Unreleased] above, and deletes consumed fragments. The --allow-empty flag ensures a no-change release still gets a dated heading (with a _No notable changes._ placeholder). A subsequent verify step confirms the promotion landed correctly. Maintainers do not run this by hand.

Archived fragments

.changeset/archived/ holds fragments for already-shipped releases (≤ 1.3.1), retained for provenance. Their content was hand-curated into the dated ## [1.x.y] sections of CHANGELOG.md during the #690 backfill — they were never consumed by render. All changeset tooling enumerates .changeset/ non-recursively, so archived fragments are never picked up or rendered. Do not move them back to the top level.