* feat: /gsd:ai-phase + /gsd:eval-review — AI evals and framework selection layer Adds a structured AI development layer to GSD with 5 new agents, 2 new commands, 2 new workflows, 2 reference files, and 1 template. Commands: - /gsd:ai-phase [N] — pre-planning AI design contract (inserts between discuss-phase and plan-phase). Orchestrates 4 agents in sequence: framework-selector → ai-researcher → domain-researcher → eval-planner. Output: AI-SPEC.md with framework decision, implementation guidance, domain expert context, and evaluation strategy. - /gsd:eval-review [N] — retroactive eval coverage audit. Scores each planned eval dimension as COVERED/PARTIAL/MISSING. Output: EVAL-REVIEW.md with 0-100 score, verdict, and remediation plan. Agents: - gsd-framework-selector: interactive decision matrix (6 questions) → scored framework recommendation for CrewAI, LlamaIndex, LangChain, LangGraph, OpenAI Agents SDK, Claude Agent SDK, AutoGen/AG2, Haystack - gsd-ai-researcher: fetches official framework docs + writes AI systems best practices (Pydantic structured outputs, async-first, prompt discipline, context window management, cost/latency budget) - gsd-domain-researcher: researches business domain and use-case context — surfaces domain expert evaluation criteria, industry failure modes, regulatory constraints, and practitioner rubric ingredients before eval-planner writes measurable criteria - gsd-eval-planner: designs evaluation strategy grounded in domain context; defaults to Arize Phoenix (tracing) + RAGAS (RAG eval) with detect-first guard for existing tooling - gsd-eval-auditor: retroactive codebase scan → scores eval coverage Integration points: - plan-phase: non-blocking nudge (step 4.5) when AI keywords detected and no AI-SPEC.md present - settings: new workflow.ai_phase toggle (default on) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: refine ai-integration-phase layer — rename, house style, consistency fixes Amends the ai-evals framework layer (df8cb6c) with post-review improvements before opening upstream PR. Rename /gsd:ai-phase → /gsd:ai-integration-phase: - Renamed commands/gsd/ai-phase.md → ai-integration-phase.md - Renamed get-shit-done/workflows/ai-phase.md → ai-integration-phase.md - Updated config key: workflow.ai_phase → workflow.ai_integration_phase - Updated repair action: addAiPhaseKey → addAiIntegrationPhaseKey - Updated all 84 cross-references across agents, workflows, templates, tests Consistency fixes (same class as PR #1380 review): - commands/gsd: objective described 3-agent chain, missing gsd-domain-researcher - workflows/ai-integration-phase: purpose tag described 3-agent chain + "locks three things" — updated to 4 agents + 4 outputs - workflows/ai-integration-phase: missing DOMAIN_MODEL resolve-model call in step 1 (domain-researcher was spawned in step 7.5 with no model variable) - workflows/ai-integration-phase: fractional step ## 7.5 renumbered to integers (steps 8–12 shifted) Agent house style (GSD meta-prompting conformance): - All 5 new agents refactored to execution_flow + step name="" structure - Role blocks compressed to 2 lines (removed verbose "Core responsibilities") - Added skills: frontmatter to all 5 agents (agent-frontmatter tests) - Added # hooks: commented pattern to file-writing agents - Added ALWAYS use Write tool anti-heredoc instruction to file-writing agents - Line reductions: ai-researcher −41%, domain-researcher −25%, eval-planner −26%, eval-auditor −25%, framework-selector −9% Test coverage (tests/ai-evals.test.cjs — 48 tests): - CONFIG: workflow.ai_integration_phase defaults and config-set/get - HEALTH: W010 warning emission and addAiIntegrationPhaseKey repair - TEMPLATE: AI-SPEC.md section completeness (10 sections) - COMMAND: ai-integration-phase + eval-review frontmatter validity - AGENTS: all 5 new agent files exist - REFERENCES: ai-evals.md + ai-frameworks.md exist and are non-empty - WORKFLOW: plan-phase nudge integration, workflow files exist + agent coverage 603/603 tests passing. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * feat: add Google ADK to framework selector and reference matrix Google ADK (released March 2025) was missing from the framework options. Adds Python + Java multi-agent framework optimised for Gemini / Vertex AI. - get-shit-done/references/ai-frameworks.md: add Google ADK profile (type, language, model support, best for, avoid if, strengths, weaknesses, eval concerns); update Quick Picks, By System Type, and By Model Commitment tables - agents/gsd-framework-selector.md: add "Google (Gemini)" to model provider interview question - agents/gsd-ai-researcher.md: add Google ADK docs URL to documentation_sources Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: adapt to upstream conventions post-rebase - Remove skills: frontmatter from all 5 new agents (upstream changed convention — skills: breaks Gemini CLI and must not be present) - Add workflow.ai_integration_phase to VALID_CONFIG_KEYS whitelist in config.cjs (config-set blocked unknown keys) - Add ai_integration_phase: true to CONFIG_DEFAULTS in core.cjs Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: rephrase 4b.1 line to avoid false-positive in prompt-injection scan "contract as a Pydantic model" matched the `act as a` pattern case-insensitively. Rephrased to "output schema using a Pydantic model". Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: adapt to upstream conventions (W016, colon refs, config docs) - Replace verify.cjs from upstream to restore W010-W015 + cmdValidateAgents, lost when rebase conflict was resolved with --theirs - Add W016 (workflow.ai_integration_phase absent) inside the config try block, avoids collision with upstream's W010 agent-installation check - Add addAiIntegrationPhaseKey repair case mirroring addNyquistKey pattern - Replace /gsd: colon format with /gsd- hyphen format across all new files (agents, workflows, templates, verify.cjs) per stale-colon-refs guard (#1748) - Add workflow.ai_integration_phase to planning-config.md reference table - Add ai_integration_phase → workflow.ai_integration_phase to NAMESPACE_MAP in config-field-docs.test.cjs so CONFIG_DEFAULTS coverage check passes - Update ai-evals tests to use W016 instead of W010 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> * fix: add 5 new agents to E2E Copilot install expected list gsd-ai-researcher, gsd-domain-researcher, gsd-eval-auditor, gsd-eval-planner, gsd-framework-selector added to the hardcoded expected agent list in copilot-install.test.cjs (#1890). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
8.8 KiB
AI-SPEC.md locks four things before the planner creates tasks:
- Framework selection (with rationale and alternatives)
- Implementation guidance (correct syntax, patterns, pitfalls from official docs)
- Domain context (practitioner rubric ingredients, failure modes, regulatory constraints)
- Evaluation strategy (dimensions, rubrics, tooling, reference dataset, guardrails)
This prevents the two most common AI development failures: choosing the wrong framework for the use case, and treating evaluation as an afterthought.
<required_reading>
@/.claude/get-shit-done/references/ai-frameworks.md
@/.claude/get-shit-done/references/ai-evals.md
</required_reading>
1. Initialize
INIT=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" init plan-phase "$PHASE")
if [[ "$INIT" == @file:* ]]; then INIT=$(cat "${INIT#@file:}"); fi
Parse JSON for: phase_dir, phase_number, phase_name, phase_slug, padded_phase, has_context, has_research, commit_docs.
File paths: state_path, roadmap_path, requirements_path, context_path.
Resolve agent models:
SELECTOR_MODEL=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" resolve-model gsd-framework-selector --raw)
RESEARCHER_MODEL=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" resolve-model gsd-ai-researcher --raw)
DOMAIN_MODEL=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" resolve-model gsd-domain-researcher --raw)
PLANNER_MODEL=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" resolve-model gsd-eval-planner --raw)
Check config:
AI_PHASE_ENABLED=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" config-get workflow.ai_integration_phase 2>/dev/null || echo "true")
If AI_PHASE_ENABLED is false:
AI phase is disabled in config. Enable via /gsd-settings.
Exit workflow.
If planning_exists is false: Error — run /gsd-new-project first.
2. Parse and Validate Phase
Extract phase number from $ARGUMENTS. If not provided, detect next unplanned phase.
PHASE_INFO=$(node "$HOME/.claude/get-shit-done/bin/gsd-tools.cjs" roadmap get-phase "${PHASE}")
If found is false: Error with available phases.
3. Check Prerequisites
If has_context is false:
No CONTEXT.md found for Phase {N}.
Recommended: run /gsd-discuss-phase {N} first to capture framework preferences.
Continuing without user decisions — framework selector will ask all questions.
Continue (non-blocking).
4. Check Existing AI-SPEC
AI_SPEC_FILE=$(ls "${PHASE_DIR}"/*-AI-SPEC.md 2>/dev/null | head -1)
If exists: Use AskUserQuestion:
- header: "Existing AI-SPEC"
- question: "AI-SPEC.md already exists for Phase {N}. What would you like to do?"
- options:
- "Update — re-run with existing as baseline"
- "View — display current AI-SPEC and exit"
- "Skip — keep current AI-SPEC and exit"
If "View": display file contents, exit. If "Skip": exit. If "Update": continue to step 5.
5. Spawn gsd-framework-selector
Display:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
GSD ► AI DESIGN CONTRACT — PHASE {N}: {name}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
◆ Step 1/4 — Framework Selection...
Spawn gsd-framework-selector with:
Read ~/.claude/agents/gsd-framework-selector.md for instructions.
<objective>
Select the right AI framework for Phase {phase_number}: {phase_name}
Goal: {phase_goal}
</objective>
<files_to_read>
{context_path if exists}
{requirements_path if exists}
</files_to_read>
<phase_context>
Phase: {phase_number} — {phase_name}
Goal: {phase_goal}
</phase_context>
Parse selector output for: primary_framework, system_type, model_provider, eval_concerns, alternative_framework.
If selector fails or returns empty: Exit with error — "Framework selection failed. Re-run /gsd-ai-integration-phase {N} or answer the framework question in /gsd-discuss-phase {N} first."
6. Initialize AI-SPEC.md
Copy template:
cp "$HOME/.claude/get-shit-done/templates/AI-SPEC.md" "${PHASE_DIR}/${PADDED_PHASE}-AI-SPEC.md"
Fill in header fields:
- Phase number and name
- System classification (from selector)
- Selected framework (from selector)
- Alternative considered (from selector)
7. Spawn gsd-ai-researcher
Display:
◆ Step 2/4 — Researching {primary_framework} docs + AI systems best practices...
Spawn gsd-ai-researcher with:
Read ~/.claude/agents/gsd-ai-researcher.md for instructions.
<objective>
Research {primary_framework} for Phase {phase_number}: {phase_name}
Write Sections 3 and 4 of AI-SPEC.md
</objective>
<files_to_read>
{ai_spec_path}
{context_path if exists}
</files_to_read>
<input>
framework: {primary_framework}
system_type: {system_type}
model_provider: {model_provider}
ai_spec_path: {ai_spec_path}
phase_context: Phase {phase_number}: {phase_name} — {phase_goal}
</input>
8. Spawn gsd-domain-researcher
Display:
◆ Step 3/4 — Researching domain context and expert evaluation criteria...
Spawn gsd-domain-researcher with:
Read ~/.claude/agents/gsd-domain-researcher.md for instructions.
<objective>
Research the business domain and expert evaluation criteria for Phase {phase_number}: {phase_name}
Write Section 1b (Domain Context) of AI-SPEC.md
</objective>
<files_to_read>
{ai_spec_path}
{context_path if exists}
{requirements_path if exists}
</files_to_read>
<input>
system_type: {system_type}
phase_name: {phase_name}
phase_goal: {phase_goal}
ai_spec_path: {ai_spec_path}
</input>
9. Spawn gsd-eval-planner
Display:
◆ Step 4/4 — Designing evaluation strategy from domain + technical context...
Spawn gsd-eval-planner with:
Read ~/.claude/agents/gsd-eval-planner.md for instructions.
<objective>
Design evaluation strategy for Phase {phase_number}: {phase_name}
Write Sections 5, 6, and 7 of AI-SPEC.md
AI-SPEC.md now contains domain context (Section 1b) — use it as your rubric starting point.
</objective>
<files_to_read>
{ai_spec_path}
{context_path if exists}
{requirements_path if exists}
</files_to_read>
<input>
system_type: {system_type}
framework: {primary_framework}
model_provider: {model_provider}
phase_name: {phase_name}
phase_goal: {phase_goal}
ai_spec_path: {ai_spec_path}
</input>
10. Validate AI-SPEC Completeness
Read the completed AI-SPEC.md. Check that:
- Section 2 has a framework name (not placeholder)
- Section 1b has at least one domain rubric ingredient (Good/Bad/Stakes)
- Section 3 has a non-empty code block (entry point pattern)
- Section 4b has a Pydantic example
- Section 5 has at least one row in the dimensions table
- Section 6 has at least one guardrail or explicit "N/A for internal tool" note
- Checklist section at end has 3+ items checked
If validation fails: Display specific missing sections. Ask user if they want to re-run the specific step or continue anyway.
11. Commit
If commit_docs is true:
git add "${AI_SPEC_FILE}"
git commit -m "docs({phase_slug}): generate AI-SPEC.md — {primary_framework} + domain context + eval strategy"
12. Display Completion
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
GSD ► AI-SPEC COMPLETE — PHASE {N}: {name}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
◆ Framework: {primary_framework}
◆ System Type: {system_type}
◆ Domain: {domain_vertical from Section 1b}
◆ Eval Dimensions: {eval_concerns}
◆ Tracing Default: Arize Phoenix (or detected existing tool)
◆ Output: {ai_spec_path}
Next step:
/gsd-plan-phase {N} — planner will consume AI-SPEC.md
<success_criteria>
- Framework selected with rationale (Section 2)
- AI-SPEC.md created from template
- Framework docs + AI best practices researched (Sections 3, 4, 4b populated)
- Domain context + expert rubric ingredients researched (Section 1b populated)
- Eval strategy grounded in domain context (Sections 5-7 populated)
- Arize Phoenix (or detected tool) set as tracing default in Section 7
- AI-SPEC.md validated (Sections 1b, 2, 3, 4b, 5, 6 all non-empty)
- Committed if commit_docs enabled
- Next step surfaced to user </success_criteria>