zavora-cli v1.1.4

Highlights

• Hardened release quality gates: fmt, clippy -D warnings, full test suite, eval quality gate, and security checks now pass on main.
• Search sub-agent behavior is now explicitly provider-scoped to Gemini and covered by regression tests.
• Improved search capability messaging with clear provider/model hinting when unavailable.
• Fixed clippy-blocking issues in runner/tool/time modules for cleaner production builds.
• Removed syntect dependency from confirmation rendering to eliminate transitive unmaintained advisories (bincode, yaml-rust) and keep cargo audit --deny warnings green.

Validation

• make release-check passed.
• Unit test suite passed (164 passed, 0 failed).
• Security checks passed (cargo audit --deny warnings + secret scanning).

Full Changelog: v1.1.1...v1.1.4

v1.1.1 — Wire 5 gaps

Fixed

1. Context usage now live — /usage shows real token breakdown from session events, prompt shows budget indicators (⚠/🔴)
2. Delegate wired — /delegate <task> runs an isolated sub-agent prompt in a separate session
3. StubTool cleaned — moved from production tool_policy.rs to test module
4. Checkpoints persisted — saved to .zavora/checkpoints.json, survives CLI restarts
5. todo_list agent tool — model can now create/complete/view/list/delete todos during execution

157 tests, 0 failures, 0 warnings.

What's Changed

• fix: wire 5 gaps — context usage, delegate, StubTool, checkpoint persistence, todo tool by @jkmaina in #60

Full Changelog: v1.1.0...v1.1.1

v1.1.0 — Phase 2: Q CLI Parity + UX

Phase 2: Q CLI Parity + UX

11 issues closed across Sprints 9–11 (PRs #49–#59).

New Features

Sprint 9 — Tool Policy and Diagnostics

• Tool aliases with wildcard allow/deny filtering (#37)
• Hook lifecycle system with 5 hook points (#38)
• MCP diagnostics with server state and auth hints (#39)

Sprint 10 — Context Management

• Context usage tracking with budget warnings (#40)
• Manual /compact and auto-compaction via ADK (#41)
• Checkpoint save/list/restore and tangent mode (#42)
• Todo list persistence and delegate experiment (#43)

Sprint 11 — UX Polish and Release

• Unified theme, command palette, fuzzy matching, onboarding (#44)
• Parity benchmark suite with 12 scenarios and scorecard (#45)
• Parity matrix (98.5%) and differentiation roadmap (#46)
• v1.1.0 GA sign-off with release gates and migration guidance (#47)

Stats

• 8 new modules: tool_policy, hooks, context, compact, checkpoint, todos, theme, benchmark
• 157 tests, 0 failures, 0 warnings
• 98.5% parity score (33/34 capabilities Met)
• No breaking changes from v1.0.0

Migration

No breaking changes. See docs/GA_SIGNOFF_v110.md for migration guidance.

What's Changed

• refactor: project optimisation — code fixes + module split by @jkmaina in #48
• feat(#37): add tool aliases and wildcard permission patterns by @jkmaina in #49
• feat(#38): implement hook lifecycle events for agent and tool execution by @jkmaina in #50
• feat(#39): expand MCP diagnostics and resilience UX by @jkmaina in #51
• feat(#40): add context usage indicator and budget warning UX by @jkmaina in #52
• feat(#41): add manual /compact and auto-compaction on overflow by @jkmaina in #53
• feat(#42): add checkpoint and tangent conversation branching by @jkmaina in #54
• feat(#43): add todo list persistence and delegate sub-agent experiment by @jkmaina in #55
• feat(#44): add unified theme, command palette, and onboarding UX by @jkmaina in #56
• feat(#45): add parity benchmark suite and scorecard by @jkmaina in #57
• docs(#46): add parity matrix and differentiation roadmap by @jkmaina in #58
• docs(#47): add v1.1.0 RC release plan and operational sign-off by @jkmaina in #59

New Contributors

• @jkmaina made their first contribution in #48

Full Changelog: v1.0.0...v1.1.0

v1.0.0

Full Changelog: v0.1.0...v1.0.0

v0.1.0

Full Changelog: https://github.com/zavora-ai/zavora-cli/commits/v0.1.0