fix(gsd): retry empty complete-slice runs

[mastertyko]

TL;DR

What: Treat empty complete-slice runs as retryable failures instead of successful completions.
Why: Context-exhausted complete-slice units could do zero successful tool calls, leave the slice pending, and block all downstream work.
How: Extend the existing zero-tool-call guard to complete-slice and add a regression that proves the loop retries instead of silently advancing.

What

This change extends the auto-loop's empty-tool-call protection so it also covers complete-slice units.

It updates src/resources/extensions/gsd/auto/phases.ts and adds regression coverage in src/resources/extensions/gsd/tests/auto-loop.test.ts.

Why

A context-exhausted complete-slice unit could finish with zero successful tool calls, leave the DB slice status unchanged, and then cycle through verification and dispatch without ever closing the slice. That silently blocks all dependent slices downstream.

Closes #2653

This PR is stacked on #3365.

How

• Reuse the existing zero-tool-call retry logic for complete-slice, not just execute-task.
• Keep the failure mode retry-oriented so the loop can recover in a fresh context instead of treating the run as success.
• Add a regression test that simulates a zero-tool-call complete-slice iteration and verifies that the loop retries before succeeding.

Change type

• fix — Bug fix
• feat — New feature or capability
• refactor — Code restructuring (no behavior change)
• test — Adding or updating tests
• docs — Documentation only
• chore — Build, CI, or tooling changes

Scope

• pi-tui — Terminal UI
• pi-ai — AI/LLM layer
• pi-agent-core — Agent orchestration
• pi-coding-agent — Coding agent
• gsd extension — GSD workflow
• native — Native bindings
• ci/build — Workflows, scripts, config

Breaking changes

• No breaking changes
• Yes — described above

Test plan

• CI passes
• New/updated tests included
• Manual testing — steps described above
• No tests needed — explained above

Verified locally with the full CI mirror gate:
npm run build
npm run typecheck:extensions
npm run test:unit
npm run test:integration

Targeted regression:
node --import ./src/resources/extensions/gsd/tests/resolve-ts.mjs --experimental-strip-types --test src/resources/extensions/gsd/tests/auto-loop.test.ts

AI disclosure

• This PR includes AI-assisted code

This PR includes AI-assisted code — prepared with Codex and verified as described in the test plan above.

[github-actions]

🔴 PR Risk Report — CRITICAL

Files changed	12
Systems affected	4
Overall risk	🔴 CRITICAL

Affected Systems

Risk	System
🔴 critical	Auto Engine
🔴 critical	State Machine
🟡 medium	Async Jobs
🟡 medium	Worktree

File Breakdown

Risk	File	Systems
🟡	src/resources/extensions/async-jobs/job-manager.ts	Async Jobs
🔴	src/resources/extensions/gsd/auto-worktree.ts	Auto Engine, Worktree
🔴	src/resources/extensions/gsd/auto/phases.ts	Auto Engine
🔴	src/resources/extensions/gsd/state.ts	State Machine
🟡	src/resources/extensions/gsd/worktree-manager.ts	Worktree
⚪	src/resources/extensions/async-jobs/await-tool.test.ts	(unclassified)
⚪	src/resources/extensions/gsd/tests/auto-loop.test.ts	(unclassified)
⚪	src/resources/extensions/gsd/tests/worktree-manager.test.ts	(unclassified)
⚪	src/resources/extensions/gsd/tests/worktree-teardown-safety.test.ts	(unclassified)
⚪	src/rtk.ts	(unclassified)
⚪	src/tests/integration/pack-install.test.ts	(unclassified)
⚪	src/tests/rtk.test.ts	(unclassified)

⚠️ Critical risk — please verify: state persistence, auth token lifecycle, agent loop race conditions, RPC protocol compatibility.

[jeremymcs]

This PR has merge conflicts with the base branch. Please rebase or merge main to resolve before review can proceed.

🤖 Automated PR audit — 2026-04-04

[mastertyko]

@jeremymcs Refreshed this PR onto the current main, reran the full local gate (npm run build, npm run typecheck:extensions, npm run test:unit, npm run test:integration), and pushed the updated branch. GitHub checks are green now.

[mastertyko]

Closing this in favor of #3300, which is the version that landed for this issue.