1be0e4e2eec8a1770e30f431bbd9b00327f4a989
7 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
|
|
8bc255c266 |
fix(workstream): normalize migration workstream names (#3269)
* fix(workstream): normalize migrate-name to valid slug * docs(context): record workstream migrate-name slug invariant * fix(catalog-cjs): balanced fallback for unknown profile (CR finding A) profiles[profile] could return undefined for any profile key absent from the catalog entry, causing downstream callers like formatAgentToModelMapAsTable to crash on .length. Add ?? profiles.balanced fallback to match the SDK adapter. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test(sdk): anchor path resolution on import.meta.url not cwd (CR finding B) resolve(process.cwd(), '..') breaks when Vitest is invoked from the repo root because cwd is already the repo root and '..' goes one level above. Replace with a file-relative path using fileURLToPath(new URL('../../../', import.meta.url)) anchored at the test file's location (sdk/src/query/). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test: derive Group B runtime list from catalog (CR finding C) Hardcoded ['kilo', 'cline', ...] throws TypeError if a runtime name is removed from the catalog. Derive group B dynamically via Object.keys(catalog.runtimeTierDefaults).filter(r => !r.opus) so the test never goes stale and auto-covers future Group B additions. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs(workflow): add hermes to Step B runtime options (CR finding D) hermes appears in the Group A built-in defaults table but was missing from the AskUserQuestion options in Step B, forcing users to manually type it via 'Other (Group B or custom)'. Add explicit hermes entry for UI consistency. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * docs(config): refresh dynamic_routing tier table; fix stale L671 (findings E+F) Finding E: tier table was missing 6 heavy-tier agents and 15 standard/light agents added by this PR. Updated all three rows to match catalog routingTier assignments (33 agents total). Finding F: removed stale '18 of 31' claim and agent enumeration; replaced with accurate note that all 33 agents have explicit catalog entries. Updated authoritative source pointers to model-catalog.cjs / model-catalog.ts. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * test(core): add profile-fallback unit tests for quality and budget (CR nitpick G) The PR introduced quality→opus and budget→haiku unknown-agent fallbacks but only balanced→sonnet and inherit→inherit were tested. Add two tests covering the remaining two branches to complete coverage. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * adr: define planning workspace and worktree seam * refactor(worktree): extract worktree safety policy module * refactor(workstream): extract active workstream pointer store seam * test(worktree): cover policy branch paths and persist seam guardrails * refactor(worktree): centralize health inventory seam for W017 * fix(workspace): align SDK project path policy with CJS planningDir * refactor(query): unify SDK planning path projection seam * refactor(init): route workspace projection through planningPaths seam * docs(adr): add SDK architecture and planning path ADRs * refactor(worktree): deepen name, pointer, inventory, and config seams * docs(config): harmonize claude-opus-4-6 to 4-7 in resolve_model_ids example (CR finding 2) * fix(sdk): return undefined for model_profile='inherit' sentinel (CR finding 3) * docs(adr): renumber conflicting 0003-sdk-package-seam-module to 0007, update seam-map reference (CR finding 4) * fix(workstream): align CJS and SDK name validation to accept dots, guard path traversal via includes('..') (CR finding 5) * fix(sdk): guard writeActiveWorkstream against non-existent workstream directory, k014/k031 parity (CR finding 6) * chore(changeset): add #3269 changeset (CR finding 1 — proper changeset for this PR) * docs(inventory): register 3 new CLI modules in INVENTORY.md/MANIFEST (active-workstream-store, workstream-name-policy, worktree-safety) * fix(sdk): use relPlanningPath(workstream) in planningPaths, fix setActiveWorkstream/getActiveWorkstream name errors in workstream.ts * fix(sdk): validate GSD_WORKSTREAM in planningPaths before use (#3269 regression) planningPaths() called resolveWorkspaceContext() which returned GSD_WORKSTREAM raw (no validation). An invalid value like '../evil' was used as effectiveWorkstream, constructing a bad path; roadmapAnalyze() caught the ENOENT and returned a no-phase_count error object instead of the root ROADMAP result. Fix: validate envCtx.workstream with validateWorkstreamName() in planningPaths() before accepting it as effectiveWorkstream. Invalid env → null → root .planning/ fallback, preserving the bug-2791 contract: invalid GSD_WORKSTREAM is silently ignored and falls back to the root context (phase_count: 0 for empty root ROADMAP). The bug-2791 regression test now passes. No other call sites read GSD_WORKSTREAM without validation: query-runtime-context.ts already validates; cli.ts already validates; context-engine.ts takes a caller-validated workstream parameter. Closes #3268 (regression introduced by #3269 workstream-name-policy work). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |
||
|
|
96806003c5 |
fix(#3229): shared model catalog source of truth for agent profiles + runtime tier defaults (#3230)
* docs(adr): add ADR-0003 model catalog module * fix(#3229): add shared model catalog as source of truth for agent profiles and runtime tier defaults Research / design (ADR-0003): - Existing drift came from 4 independent model truths: 1. CJS model-profiles.cjs 2. SDK config-query.ts stale copy (18 agents) 3. settings-advanced.md runtime tier table 4. session-runner Claude-only profile map - New design: one machine-readable Model Catalog Module in sdk/shared/ that both packages ship and consume. Implementation: - sdk/shared/model-catalog.json — canonical source of truth for: - full 33-agent registry - per-agent golden (quality) alias + balanced/budget aliases - adaptive derivation from routingTier - agent→phaseType map - agent→dynamic-routing default tier map - runtime tier defaults for all supported runtimes - get-shit-done/bin/lib/model-catalog.cjs — CJS adapter over the catalog - sdk/src/model-catalog.ts — SDK adapter over the same catalog - CJS model-profiles.cjs now re-exports derived data from model-catalog.cjs - SDK config-query.ts now re-exports MODEL_PROFILES/VALID_PROFILES from model-catalog.ts instead of maintaining its own list - sdk/src/query/helpers.ts runtime list now comes from the catalog (fixes hermes drift) - sdk/src/session-runner.ts Claude profile→model-id mapping now resolves via catalog - docs/CONFIGURATION.md + settings-advanced.md runtime tables updated to match catalog Behavior changes: - resolve-model now covers every shipped agent file on disk (33 agents) - unknown-agent fallback is profile-semantic, not hardcoded sonnet: quality→opus, budget→haiku, balanced/adaptive→sonnet, inherit→inherit - Group B runtimes remain known runtimes but do not get built-in tier defaults Tests (RED→GREEN): - root tests: shipped agent files must equal MODEL_PROFILES keys - sdk tests: shipped agent files must equal MODEL_PROFILES keys - direct fix assertion: gsd-code-reviewer resolves to opus under quality with no unknown_agent - runtime defaults parity test: settings-advanced.md + CONFIGURATION.md tables must match catalog - helper tests: hermes included in SUPPORTED_RUNTIMES and getRuntimeConfigDir() Closes #3229 * chore(changeset): update #3229 changeset pr field to 3230 * fix(ci): update inherit fallback expectations and inventory parity for model catalog |
||
|
|
f9ed47ac8b |
fix(#2832): gsd-sdk auto detects Codex runtime correctly (#2844)
* fix(#2832): gsd-sdk auto detects Codex runtime correctly Two-part fix for #2832 (gsd-sdk auto silently routing non-Claude runtime projects through the Claude Agent SDK): 1. Runtime gate at the `auto` entry point. New `runtime-gate.ts` exports `assertRuntimeSupportsAutoMode(config)` which throws an actionable error when `GSD_RUNTIME` / `config.runtime` resolves to a non-Claude runtime (codex, gemini, opencode, etc.). The autonomous orchestrator only knows how to drive `@anthropic-ai/claude-agent-sdk` today; failing fast with a clear pointer at the in-session slash commands beats the previous instant `[FAILED] $0.00 0.1s` flake. Wired into `cli.ts` before the GSD/InitRunner construction. 2. Runtime-aware `resolveModel()` in `session-runner.ts`. The profile -> id map (`balanced -> claude-sonnet-4-6`, etc.) was applied unconditionally, so even with `runtime: codex` and `resolve_model_ids: omit` the SDK forced a Claude id into `query()`. Now the profile map only fires when the runtime is Claude and the explicit `resolve_model_ids: "omit"` knob short-circuits to undefined, mirroring `query/config-query.ts`. Tests (vitest, sdk/src): - runtime-gate.test.ts (8 cases): claude / unset / unknown pass; codex, gemini, opencode throw; GSD_RUNTIME wins over config.runtime; error message references #2832 and the slash-command workaround. - session-runner.test.ts (4 new cases under "resolveModel runtime awareness (#2832)"): codex runtime + balanced profile -> no model injected; resolve_model_ids: omit -> no model; claude runtime still resolves to claude-sonnet-4-6 (no regression); explicit options.model wins on any runtime. * fix(#2832): address CR — env-precedence in resolveModel + accurate source attribution Two CodeRabbit findings on PR #2844: 1. session-runner.ts:resolveModel() (Major) — read runtime via detectRuntime() so GSD_RUNTIME env precedence is honored. Without this, a Codex run with a Claude-shaped config still fell into the Claude-only profile-id branch. 2. runtime-gate.ts:assertRuntimeSupportsAutoMode() (Minor) — when GSD_RUNTIME holds an unsupported value, detectRuntime() falls through to config but the source label still reported the discarded env value. Fix: validate env against SUPPORTED_RUNTIMES before attributing the source. Tests added for both: env-precedence in session-runner, source attribution in runtime-gate. 17/17 pass. |
||
|
|
2fafbd2753 |
fix(sdk): pass phaseDir to executor prompt so SUMMARY.md lands in .planning/
The SDK's buildExecutorPrompt told executors to "Create a SUMMARY.md file"
with no directory path, causing them to write it in cwd (project root)
instead of .planning/phases/{phase}/. Thread phaseDir from PhaseRunner
through PromptFactory and into the completion instructions so the executor
gets an explicit path like `.planning/phases/01-auth/01-01-SUMMARY.md`.
Backward compatible — buildExecutorPrompt still accepts a plain string
(agentDef) for existing callers.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
||
|
|
1005f02db2 |
fix(sdk): stop duplicating prompt in runPhaseStepSession user message (#2223)
runPhaseStepSession was passing the full prompt string as both the user-visible prompt: argument and as systemPrompt.append, sending the same (potentially large) text twice per invocation and doubling the token cost for every phase step session. runPlanSession correctly uses a short directive as the user message and reserves the full content for systemPrompt.append only. Apply the same pattern to runPhaseStepSession: use a brief "Execute this phase step: <step>" directive as the user message. Closes #2194 Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com> |
||
|
|
ade67cf9f9 |
fix: update MODEL_ALIAS_MAP to current Claude model IDs (#1691)
Fixes #1690 - opus: claude-opus-4-0 → claude-opus-4-6 - sonnet: claude-sonnet-4-5 → claude-sonnet-4-6 - haiku: claude-haiku-3-5 → claude-haiku-4-5 Also updates the stale haiku reference in sdk/src/session-runner.ts and documentation examples in CONFIGURATION.md (en, ja-JP, ko-KR). |
||
|
|
596ce2d252 |
feat: GSD SDK — headless CLI with init + auto commands (#1407)
* test: Bootstrapped sdk/ as TypeScript ESM package with full GSD-1 PLAN.… - "sdk/package.json" - "sdk/tsconfig.json" - "sdk/vitest.config.ts" - "sdk/src/types.ts" - "sdk/src/plan-parser.ts" - "sdk/src/plan-parser.test.ts" GSD-Task: S01/T01 * test: Implemented config reader and gsd-tools bridge with 25 unit tests… - "sdk/src/config.ts" - "sdk/src/config.test.ts" - "sdk/src/gsd-tools.ts" - "sdk/src/gsd-tools.test.ts" GSD-Task: S01/T02 * test: Built prompt-builder, session-runner, and GSD class — 85 total un… - "sdk/src/prompt-builder.ts" - "sdk/src/prompt-builder.test.ts" - "sdk/src/session-runner.ts" - "sdk/src/index.ts" - "sdk/src/types.ts" GSD-Task: S01/T03 * test: Created E2E integration test with fixtures proving full SDK pipel… - "sdk/src/e2e.integration.test.ts" - "sdk/test-fixtures/sample-plan.md" - "sdk/test-fixtures/.planning/config.json" - "sdk/test-fixtures/.planning/STATE.md" - "vitest.config.ts" - "tsconfig.json" GSD-Task: S01/T04 * test: Added PhaseType/GSDEventType enums, 16-variant GSDEvent union, GS… - "sdk/src/types.ts" - "sdk/src/event-stream.ts" - "sdk/src/logger.ts" - "sdk/src/event-stream.test.ts" - "sdk/src/logger.test.ts" GSD-Task: S02/T01 * test: Built ContextEngine for phase-aware context file resolution, getT… - "sdk/src/context-engine.ts" - "sdk/src/tool-scoping.ts" - "sdk/src/phase-prompt.ts" - "sdk/src/context-engine.test.ts" - "sdk/src/tool-scoping.test.ts" - "sdk/src/phase-prompt.test.ts" GSD-Task: S02/T02 * test: Wired event stream into session runner, added onEvent()/addTransp… - "sdk/src/session-runner.ts" - "sdk/src/index.ts" - "sdk/src/e2e.integration.test.ts" GSD-Task: S02/T03 * feat: Added PhaseStepType enum, PhaseOpInfo interface, phase lifecycle… - "sdk/src/types.ts" - "sdk/src/gsd-tools.ts" - "sdk/src/session-runner.ts" - "sdk/src/index.ts" - "sdk/src/phase-runner-types.test.ts" GSD-Task: S03/T01 * test: Implemented PhaseRunner state machine with 39 unit tests covering… - "sdk/src/phase-runner.ts" - "sdk/src/phase-runner.test.ts" GSD-Task: S03/T02 * test: Wired PhaseRunner into GSD.runPhase() public API with full re-exp… - "sdk/src/index.ts" - "sdk/src/phase-runner.integration.test.ts" - "sdk/src/phase-runner.ts" GSD-Task: S03/T03 * test: Expanded runVerifyStep with full gap closure cycle (plan → execut… - "sdk/src/types.ts" - "sdk/src/phase-runner.ts" - "sdk/src/phase-runner.test.ts" GSD-Task: S04/T02 * fix: Added 3 integration tests proving phasePlanIndex returns correct t… - "sdk/src/phase-runner.integration.test.ts" - "sdk/src/index.ts" GSD-Task: S04/T03 * test: Add milestone-level types, typed roadmapAnalyze(), GSD.run() orch… - "sdk/src/types.ts" - "sdk/src/gsd-tools.ts" - "sdk/src/index.ts" - "sdk/src/milestone-runner.test.ts" GSD-Task: S05/T01 * test: Added CLITransport (structured stdout log lines) and WSTransport… - "sdk/src/cli-transport.ts" - "sdk/src/cli-transport.test.ts" - "sdk/src/ws-transport.ts" - "sdk/src/ws-transport.test.ts" - "sdk/src/index.ts" - "sdk/package.json" GSD-Task: S05/T02 * test: Added gsd-sdk CLI entry point with argument parsing, bin field, p… - "sdk/src/cli.ts" - "sdk/src/cli.test.ts" - "sdk/package.json" GSD-Task: S05/T03 * feat: Add InitNewProjectInfo type, initNewProject()/configSet() GSDTool… - "sdk/src/types.ts" - "sdk/src/gsd-tools.ts" - "sdk/src/cli.ts" - "sdk/src/cli.test.ts" - "sdk/src/gsd-tools.test.ts" GSD-Task: S01/T01 * chore: Created InitRunner orchestrator with setup → config → PROJECT.md… - "sdk/src/init-runner.ts" - "sdk/src/types.ts" - "sdk/src/index.ts" GSD-Task: S01/T02 * test: Wired InitRunner into CLI main() for full gsd-sdk init dispatch a… - "sdk/src/cli.ts" - "sdk/src/init-runner.test.ts" - "sdk/src/cli.test.ts" GSD-Task: S01/T03 * test: Add PlanCheck step, AI self-discuss, and retryOnce wrapper to Pha… - "sdk/src/types.ts" - "sdk/src/phase-runner.ts" - "sdk/src/session-runner.ts" - "sdk/src/phase-runner.test.ts" - "sdk/src/phase-runner-types.test.ts" GSD-Task: S02/T01 * feat: Rewrite CLITransport with ANSI colors, phase banners, spawn indic… - "sdk/src/cli-transport.ts" - "sdk/src/cli-transport.test.ts" GSD-Task: S02/T02 * test: Add `gsd-sdk auto` command with autoMode config override, USAGE t… - "sdk/src/cli.ts" - "sdk/src/cli.test.ts" - "sdk/src/index.ts" - "sdk/src/types.ts" GSD-Task: S02/T03 * fix: CLI shebang + gsd-tools non-JSON output handling Three bugs found during first real gsd-sdk run: 1. cli.ts shebang was commented out — shell executed JS as bash, triggering ImageMagick's import command instead of Node 2. configSet() called exec() which JSON.parse()d the output, but gsd-tools config-set returns 'key=value' text, not JSON. Added execRaw() method for commands that return plain text. 3. Same JSON parse bug affected commit() (returns git SHA), stateLoad(), verifySummary(), initExecutePhase(), stateBeginPhase(), and phaseComplete(). All switched to execRaw(). Tests updated to match real gsd-tools output format (plain text instead of mocked JSON). 376/376 tests pass. |