Phonton CLI v0.20.1

phonton-cli v0.20.1

New

• phonton serve — local JSON-RPC sidecar for Phonton Desktop (POST /rpc, SSE /events/<task_id>).
• HandoffPacket.schema_version — stable receipt export contract ("1").

Methods (serve)

• ping, plan.preview, doctor.run, review.get, goal.start, goal.status

Ship checklist

1. RunIndex 38 benchmark evidence on disk (C:\benchmark-phonton)
2. cargo test --locked --workspace from clean clone
3. Tag v0.20.1 from phonton-dev mirror → public phonton-cli clone

phonton-cli v0.19.7

v0.19.7 - Local Template Dispatch And Launch Benchmark Harness

Summary

Market-ready reliability release: wires embedded worker templates back into the
execution path, fixes the chess rules test seed for Vitest-free verification,
and adds the memory-latency benchmark test hook.

Highlights

• Local template dispatch for chess rules seed, syntax preflight, and receipt refactor slices.
• PHONTON_DISABLE_LOCAL_SEEDS=1 for provider-only benchmark runs.
• benchmark_latency_concurrent memory harness test.

Validation

cargo fmt --all -- --check
cargo test --locked --workspace
cargo clippy --locked --workspace --all-targets -- -D warnings
cargo build --release --locked -p phonton-cli

Benchmark workspace:

cd C:\benchmark-phonton
powershell -NoProfile -ExecutionPolicy Bypass -File .\dev\Invoke-PhontonCliTest.ps1 -CliRepo C:\phonton\_phonton-cli-v0.19.0 -Build Release -Mode Smoke
powershell -NoProfile -ExecutionPolicy Bypass -File .\dev\Invoke-PhontonBenchmarks.ps1 -CliRepo C:\phonton\_phonton-cli-v0.19.0 -Build Release -RunIndex 33

Claims

Do not publish token-leaderboard copy from local-template runs. Use only
token_claim_eligible provider-reported artifacts for comparison claims.

Phonton CLI v0.19.5

v0.19.5 — Confidence Gate for Under-Specified Goals

Welcome to the v0.19.5 release of Phonton CLI!

This release introduces an essential cost-safety safeguard: the Goal Contract Confidence Gate.

🚀 Key Improvements

1. 🛡️ Goal Contract Confidence Gate

• The Problem: Broad or vague stackless goals (e.g. "make chess") were previously executed blindly by the subtask scheduler. This resulted in workers writing poor, dummy 15-line stubs (like an empty 2D board array) while wasting 12,456+ tokens across retry attempts and context compilation.
• The Fix: We've introduced a strict check inside phonton-orchestrator's scheduling intake:
  • If the goal's generated GoalContract has a confidence below 70% (e.g. <= 3 words in description) OR has unanswered clarification_questions (e.g. <= 2 words in description), the task is immediately halted.
  • The orchestrator emits a TaskFailed event, sets the state to Failed, and populates the handoff packet gaps.
  • The TUI defaults focus to Problems view, cleanly presenting the required clarification questions so users can refine their goals without wasting a single token.

---

🛠️ Verification & Quality Gates

• Formatting (rustfmt): Fully verified and automated.
• Crate Compiler & Lints (clippy): Passed with zero warnings or errors.
• Unit & Integration Tests: 100% pass rate including the new confidence_gate_halts_low_confidence_goal test suite.

Phonton CLI v0.19.4

v0.19.4 — Double-Line Splash Logo & RAII Browser Verifier Polish

Added

• RAII Browser Check Guard: Implemented a robust BrowserCheckGuard helper struct in phonton-verify/src/browser.rs using the RAII (Resource Acquisition Is Initialization) pattern. This guarantees that all temporary files (phonton-server.js and phonton-playwright.js) are deleted and spawned background Node processes are cleanly terminated across all possible exit paths (errors, success, or early returns). This completely eliminates zombie Node process leaks on the host system.

Fixed

• Double-Line Wordmark Reversion: Reverted the TUI splash logo to the classic 7-line box-drawing shadow glyph wordmark (██████╗ / ╚═╝) to preserve classic product branding.
• Gradient Animation Removal: Set the TUI shimmer phase to a static 0.0 inside render_splash. This completely removes animation refresh overlaps and line-height jitter while preserving the premium color gradient.

Phonton CLI v0.19.3

v0.19.3 — Solid-Block TUI Splash & Silent Embedder Downloads

Fixed

• Logo Alignment: Replaced double-line box-drawing characters in the TUI splash logo with a robust, solid block-character wordmark. This fixes visual line overlaps, vertical misalignments, and rendering bugs across different standard monospaced terminal fonts.
• Silent Sentence Embedder Downloads: Added strict suppression to standard output progress indicators when loading the sentence-embedding ONNX model. This prevents the fastembed telemetry from corrupting the TUI input line and rendering layout during background downloads.

Phonton CLI v0.19.2

v0.19.2 — TUI Version & Auto-Update Engine

Added

• TUI Version Number: Shows the active Phonton CLI version inside the sidebar's "System" panel.
• Auto-Update Engine: Checks for updates asynchronously at startup via the npm registry. Mirrors the Gemini CLI settings schema:
  • Enabled by default.
  • Can be turned off by setting general.enable_auto_update = false (or enableAutoUpdate = false) in your config.toml.
  • On TUI exit, if a new version is found, Phonton launches a global detached update process (npm install -g phonton-cli@latest) with a built-in safety sleep to prevent Windows file-locking failures.

Phonton CLI v0.19.1

v0.19.1 — Shadow Logo Restore

Fixed

• Restored the standard ANSI Shadow Phonton ASCII splash logo with the full 7-line block-border wordmark and ░▒▓█...▓▒░ glow strip.
• Upgraded the logo_line shader to dual-wave shimmer with separate styling for block borders (╔═╗║╝╚), shadow glyphs (░▒▓), and solid fills.

Phonton CLI v0.19.0

Phonton CLI v0.19.0

v0.19.0 ships an alpha slice of Phonton's next ADE architecture: typed swarm
planning, selectable code-index backends, and preview-only MCP capability
negotiation. It keeps the current product loop intact:

goal -> plan -> edit -> verify -> review -> remember

Added

• PlanGraph and SubtaskAssignment metadata in planner output, including
  role, expected touch scope, dependencies, conflict group, and verification
  intent.
• Broad-goal swarm activation with provider-backed planning when configured and
  visible deterministic fallback reasons when it is not.
• Conflict-aware dependency insertion in the orchestrator so overlapping touch
  scopes do not race while isolated subtasks still run concurrently.
• A CodeRetriever abstraction with default local-hnsw retrieval and an
  optional Qdrant HTTP backend for code and symbol retrieval.
• [index] config:

[index]
backend = "local-hnsw"

[index]
backend = "qdrant"
qdrant_url = "http://127.0.0.1:6333"
qdrant_collection = "phonton-code"

• phonton mcp capabilities <server-id> [--json] [--yes], which previews MCP
  initialize metadata, tool descriptors, and proposed permission rules without
  calling tools or writing config.

Changed

• phonton plan --json now includes swarm/index evidence fields.
• phonton doctor reports the selected index backend and checks Qdrant
  connectivity when Qdrant is configured.
• Flight Log telemetry includes MCP capability discovery evidence.
• Context and outcome evidence can carry plan graph and selected index backend
  data for review and future resume surfaces.

Guardrails

• Qdrant is an optional external code retrieval backend. Phonton does not start
  containers, manage collections, or migrate decision memory into Qdrant.
• SQLite keyword memory remains the authoritative local decision memory store.
• MCP capability negotiation is preview-only. It proposes sandbox permission
  rules for review but does not silently grant permissions or write config.
• v0.19.0 does not claim measured performance wins. Public benchmark claims
  still require reproducible fixtures, exact prompts, tool versions, provider
  token usage where available, raw logs, final diffs, verification logs, and
  handoff evidence.

Validation Checklist

cargo fmt --all -- --check
cargo test --locked --workspace
cargo clippy --locked --workspace --all-targets -- -D warnings
cargo build --release --locked -p phonton-cli
target/release/phonton.exe version
target/release/phonton.exe --help
npm run test:npm-wrapper
npm pack --dry-run

Phonton CLI v0.18.0

Automated Phonton CLI release assets.

Phonton CLI v0.17.1

Phonton CLI v0.17.1

DeepSeek-TUI Benchmark Enrolment & HNSW Semantic Recall

Phonton CLI v0.17.1 officially integrates and supports Hunter Bown's DeepSeek-TUI directly in the benchmark setup and capture harness. This release also marks the stabilization of our dynamic provider key mapping, credentials autopilot system, and high-speed local HNSW vector memory retrieval.

---

What's New in v0.17.1

1. DeepSeek-TUI Benchmark Enrolment

• Officially integrates and supports Hunter Bown's DeepSeek-TUI (v0.8.39) across all benchmark suites.
• Achieves a clean, 0-error validation score on comparative runs.

2. Dynamic Keys & Credentials Autopilot

• Dynamic Key Map: Allows configuring side-by-side multi-provider keys inside config.toml under the [provider.keys] section.
• Dynamic Key Resolver: Dynamically resolves {PROVIDER}_API_KEY environment variables (e.g. KIMI_API_KEY, DEEPSEEK_API_KEY, GROK_API_KEY) and maps provider tiers to defaults.
• Credentials Autopilot: Command phonton providers import-opencode automatically scans standard app folders, parses credentials, and safely merges them into global configurations.

3. High-Speed HNSW Semantic Memory

• Employs local cosine vector similarity indexing (usearch) and lightweight, local ONNX embeddings (fastembed with all-MiniLM-L6-v2).
• Performs workspace semantic recall queries in under 160µs query latency.

4. AST Syntax Preflight Checks

• Brand new 05-syntax-preflight benchmark suite verifying that Layer 1 tree-sitter AST validation catches syntax bugs across Rust, Python, and TypeScript changed files.