SDK feature parity: 29 features need active exercise testing

[bradygaster]

Summary

Systematic SDK-first feature parity testing in the SdkSquad test bed has identified 29 features that need active exercise testing. These features have correct filesystem setup (configs, templates, routing rules exist) but their runtime behavior hasn't been verified for SDK squads.

This issue tracks the remaining test matrix. Each feature category below needs a dedicated test session.

Test Matrix — Remaining Features

P1: Workflow Features (7 features)

#	Feature	What to Test
19	Directive capture	Give "Always..." directive → verify inbox capture
20	Eager execution	Multi-domain task → verify anticipatory spawns
22	GitHub repo connection	Connect to repo → verify Issue Source in team.md
23	Issue triage	Label issue squad → verify Lead assigns squad:{member}
24	Issue lifecycle	Issue → branch → PR → merge flow
25	Label routing	Apply squad:{name} → verify agent dispatch
29-32	Ralph (4 features)	Activation, work-check, idle-watch, watch mode

P2: Advanced Features (12 features)

#	Feature	What to Test
26-28	Ceremonies (3)	Auto-trigger, manual, cooldown
33-35	PRD Mode (3)	Intake, decomposition, work routing
36-38	Human & Copilot (3)	Add human, add @copilot, auto-assign
39-40	Skills (2)	Agent extraction, confidence lifecycle

P3: Platform & Edge Cases (9 features)

#	Feature	What to Test
42	Worktree awareness	Multi-worktree team root resolution
45-46	Reviewer protocol (2)	Rejection lockout, deadlock escalation
47	Client compatibility	VS Code mode detection & degradation
48	MCP integration	Tool discovery, graceful degradation
49	Constraint budget	Counter display during constrained tasks
50	Multi-agent artifact	Assembled result + raw appendix format

Already Verified (this session)

#	Feature	Status
17	Skill-aware routing	✅ Verified — skills/ exists with correct format
44	Merge drivers	✅ Verified — .gitattributes has merge=union rules
33	PRD intake	✅ Partially verified — PRD.md exists, referenced in decisions

Blocking Issues

• SDK init: Adding team members doesn't update squad.config.ts #337 — Config ↔ team.md sync (blocks Feature: Get squads to use project boards #6, Please Don't Lock squad into using GitHub #8, Copilot client parity: Squad works fully only on GitHub Copilot CLI #10)
• Ralph missing from config (blocks New team members should leverage configured plugin marketplace #29–VS Code: Test and document runSubagent as Squad spawn mechanism #32) — see companion issue
• @copilot routing without roster (blocks Context window optimization: decisions.md pruning, spawn template dedup, Init Mode compression #37–docs: MCP, notifications, and plugin marketplace guides #38) — see companion issue

Test Bed

• Repo: SdkSquad (Snake game, vanilla JS)
• PRD: PRD.md at repo root tracks all 50 features
• Squad: Matrix universe (Neo, Tank, Switch, Dozer, Scribe, Ralph)

Exit Criteria

Every feature is either:

• ✅ Verified — tested and confirmed working in SDK squad
• 🚫 Scoped out — explicitly excluded with rationale

No ⚠️ Untested or ❌ Known Gap items remain.

[github-actions]

🏗️ Squad Triage — Flight (Lead)

Issue: #340 — SDK feature parity: 29 features need active exercise testing
Assigned to: FIDO (Quality Owner)
Reason: Issue relates to testing/quality work

@copilot evaluation: 🔴 Not suitable — issue involves work outside the coding agent's capability profile.

Team roster:

• Flight (Lead) → label: squad:flight
• Procedures (Prompt Engineer) → label: squad:procedures
• EECOM (Core Dev) → label: squad:eecom
• FIDO (Quality Owner) → label: squad:fido
• PAO (DevRel) → label: squad:pao
• CAPCOM (SDK Expert) → label: squad:capcom
• CONTROL (TypeScript Engineer) → label: squad:control
• Surgeon (Release Manager) → label: squad:surgeon
• Booster (CI/CD Engineer) → label: squad:booster
• GNC (Node.js Runtime) → label: squad:gnc
• Network (Distribution) → label: squad:network
• RETRO (Security) → label: squad:retro
• INCO (CLI UX & Visual Design) → label: squad:inco
• GUIDO (VS Code Extension) → label: squad:guido
• Telemetry (Aspire & Observability) → label: squad:telemetry
• VOX (REPL & Interactive Shell) → label: squad:vox
• DSKY (TUI Engineer) → label: squad:dsky
• Sims (E2E Test Engineer) → label: squad:sims
• Handbook (SDK Usability) → label: squad:handbook
• Ralph (Work Monitor) → label: squad:ralph
• @copilot (Coding Agent) → label: squad:copilot

To reassign, remove the current squad:* label and add the correct one.

[bradygaster]

Update: Testing Progress (2026-03-10)

Original filing said 29 features untested. After systematic testing this session:

Before: 18 ✅ | 3 ❌ | 29 ⚠️
After: 32 ✅ | 6 ❌ | 12 ⚠️

Features verified this session (+14):

• Add / Update documentation on Ralph setup - PAT Classic needed #17 Skill-aware routing ✅
• Display Squad version in agent name across Copilot hosts #19 Directive capture ✅
• Community page: track contributors, stargazers, forks, and link from README #20 Eager execution ✅
• Agent progress updates: surface periodic status during long-running work #22 GitHub repo connection ✅
• Self-testing framework: observe UX patterns and identify self-improvement opportunities #23 Issue triage ✅
• Research: Run Squad from the Copilot Coding Agent (CCA) #25 Label routing ✅
• Explore Work IQ + Squad integration opportunities #26 Auto-triggered ceremonies ✅
• New team members should leverage configured plugin marketplace #29 Ralph activation ✅
• npx github:bradygaster/squad appears to hang when no SSH agent is running #30 Work-check cycle ✅
• VS Code: Test agent file discovery and .ai-team/ access #33 PRD intake ✅
• VS Code: Model selection and background mode parity #34 Lead decomposition ✅
• Create client parity compatibility matrix document #35 Work item routing ✅
• test: MCP config and plugin marketplace tests #39 Skill extraction ✅
• ci: add main branch guard workflow #44 Merge drivers ✅
• chore: trim npm package — exclude docs and contributor files #48 MCP integration ✅

Remaining 12 features (need specific test setups):

#	Feature	What's needed
24	Issue→PR lifecycle	Real issue pickup with branch/PR/merge
27	Manual ceremonies	User-requested ceremony during multi-agent work
28	Ceremony cooldown	Two ceremonies in sequence
31	Ralph idle-watch	Long-running session with periodic poll
32	Watch mode	npx squad watch exercise
36	Human members	Add human + test coordinator pause
40	Skill confidence	Independent validation to bump low→medium
42	Worktree awareness	git worktree add for multi-worktree
45	Reviewer lockout	Review/rejection scenario
46	Deadlock handling	All-agents-locked escalation
47	Client compatibility	VS Code Copilot Chat testing
49	Constraint budget	Active constraint session
50	Multi-agent artifact	Multi-agent deliverable assembly

New gaps found (+3):

• SDK init: Ralph (Work Monitor) missing from generated squad.config.ts #338 — Ralph missing from config
• SDK init: @copilot routing configured without team.md roster entry #339 — @copilot routing without roster

Full PRD with complete matrix: #341