pythinker-code 0.19.0__tar.gz → 0.21.0__tar.gz

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pythinker-code
-Version: 0.19.0
+Version: 0.21.0
 Summary: Pythinker Code is your next CLI agent.
 Keywords: cli,agent,ai,coding-assistant,llm,claude,openai,terminal
 Author: Mohamed Elkholy
@@ -79,7 +79,7 @@ Description-Content-Type: text/markdown
 [![License: Apache-2.0](https://img.shields.io/badge/License-Apache--2.0-16a34a.svg?style=for-the-badge)](https://github.com/mohamed-elkholy95/Pythinker-Code/blob/main/LICENSE)
 [![CI](https://img.shields.io/github/actions/workflow/status/mohamed-elkholy95/Pythinker-Code/ci-pythinker-cli.yml?branch=main&label=CI&style=for-the-badge&logo=githubactions&logoColor=white)](https://github.com/mohamed-elkholy95/Pythinker-Code/actions/workflows/ci-pythinker-cli.yml?query=branch%3Amain)
 
-[![Downloads](https://img.shields.io/pepy/dt/pythinker-code?cacheSeconds=300)](https://pepy.tech/projects/pythinker-code)
+[![PyPI downloads](https://assets.piptrends.com/get-last-month-downloads-badge/pythinker-code.svg)](https://piptrends.com/package/pythinker-code)
 [![Code style: Ruff](https://img.shields.io/badge/code%20style-ruff-f59e0b.svg?style=flat-square&logo=ruff&logoColor=white)](https://docs.astral.sh/ruff/)
 [![ACP ready](https://img.shields.io/badge/ACP-ready-7c3aed.svg?style=flat-square)](https://github.com/agentclientprotocol/agent-client-protocol)
 [![MCP tools](https://img.shields.io/badge/MCP-tools-0891b2.svg?style=flat-square)](https://modelcontextprotocol.io/)
@@ -115,14 +115,11 @@ It speaks the [**Agent Client Protocol (ACP)**](https://github.com/agentclientpr
 
 ---
 
-## 🆕 What's New in 0.19.0
+## 🆕 What's New in 0.21.0
 
-- **Central per-project agent memory.** Pythinker now keeps a durable per-project memory (`MEMORY.md` / `USER.md`) under `~/.pythinker/projects/<key>/memory/`, written through a new root-only `Memory` tool with content guards and secret-shape detection, recalled into the root agent's first wakeup prompt within a bounded budget, and inspectable with the new `/memory` command.
-- **Non-blocking update flow.** The blocking pre-start update prompt is replaced by a cached, no-network startup notice plus a triggerable `/update` command, and the native Windows installer now waits on the launching process before swapping files and cleans up its staged installer.
-- **Fully native Homebrew formula.** Brew installs are generated from the same GitHub Release tarballs as the curl installer (including macOS Intel native assets), so they no longer depend on PyPI virtualenv resources.
-- **Release pipeline hardening.** TestPyPI publishing is now non-blocking and publish steps carry timeouts, so a transient staging flake no longer fails a release.
+- **Non-blocking memory recall.** Recall I/O is now dispatched to a thread-pool executor, preventing the event loop from stalling during memory lookups. Sessions with large memory stores or slow disks will remain responsive throughout recall.
 
-Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.19.0`, or use the native installer for your platform from the [Releases page](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest).
+Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.21.0`, or use the native installer for your platform from the [Releases page](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest).
 
 
 ---
@@ -212,7 +209,7 @@ matches your OS — no Python, Node, or `uv` prerequisite.
 
 | Platform | Recommended install | Artifact source |
 |---|---|---|
-| **🪟 Windows** | `irm https://pythinker.com/install.ps1 \| iex` | `PythinkerSetup-0.19.0.exe` from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
+| **🪟 Windows** | `irm https://pythinker.com/install.ps1 \| iex` | `PythinkerSetup-0.21.0.exe` from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
 | **<img src="https://img.shields.io/badge/-macOS-000000?style=flat-square&logo=apple&logoColor=white" alt="macOS"> / <img src="https://img.shields.io/badge/-Linux-FCC624?style=flat-square&logo=linux&logoColor=black" alt="Linux">** | `curl -fsSL https://pythinker.com/install.sh \| bash` | native tarball from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
 | **<img src="https://img.shields.io/badge/-macOS-000000?style=flat-square&logo=apple&logoColor=white" alt="macOS"> — Homebrew** | `brew install mohamed-elkholy95/pythinker/pythinker-code` | auto-published Homebrew tap |
 | **<img src="https://img.shields.io/badge/-Linux-FCC624?style=flat-square&logo=linux&logoColor=black" alt="Linux"> — system package** | Download the `.deb` or `.rpm` for your distro below | [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
@@ -237,7 +234,7 @@ pythinker                      # start the interactive TUI
 
 ### 🪟 Windows — native installer
 
-`PythinkerSetup-0.19.0.exe` is a signed* Inno Setup wizard. Installs per-user
+`PythinkerSetup-0.21.0.exe` is a signed* Inno Setup wizard. Installs per-user
 into `%LOCALAPPDATA%\Programs\Pythinker`, registers `pythinker` on your user
 PATH (`HKCU\Environment`), broadcasts `WM_SETTINGCHANGE` so new shells see
 the change. **No UAC prompt.**
@@ -248,13 +245,13 @@ irm https://pythinker.com/install.ps1 | iex
 
 # Or manually download the installer + checksum from the Releases page,
 # verify with Get-FileHash, then run:
-.\PythinkerSetup-0.19.0.exe
+.\PythinkerSetup-0.21.0.exe
 
 # Open a fresh PowerShell
 pythinker --version
 ```
 
-**Per-machine install** (IT-managed boxes): `.\PythinkerSetup-0.19.0.exe /ALLUSERS`
+**Per-machine install** (IT-managed boxes): `.\PythinkerSetup-0.21.0.exe /ALLUSERS`
 installs to `%ProgramFiles%\Pythinker` and writes PATH to HKLM (requires admin).
 
 **Upgrade:** `pythinker update` from inside the running app — it downloads
@@ -305,26 +302,26 @@ attached to every GitHub Release.
 
 ```sh
 # Debian / Ubuntu (x86_64)
-sudo dpkg -i pythinker-code_0.19.0_amd64.deb
+sudo dpkg -i pythinker-code_0.21.0_amd64.deb
 sudo apt-get install -f       # only if dpkg reports missing deps
 
 # Debian / Ubuntu (ARM64)
-sudo dpkg -i pythinker-code_0.19.0_arm64.deb
+sudo dpkg -i pythinker-code_0.21.0_arm64.deb
 
 # Fedora / RHEL / openSUSE (x86_64)
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.x86_64.rpm
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.x86_64.rpm.sha256
-sha256sum -c pythinker-code-0.19.0.x86_64.rpm.sha256
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.x86_64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.x86_64.rpm.sha256
+sha256sum -c pythinker-code-0.21.0.x86_64.rpm.sha256
 # Fedora / RHEL:
-sudo dnf install ./pythinker-code-0.19.0.x86_64.rpm
+sudo dnf install ./pythinker-code-0.21.0.x86_64.rpm
 # openSUSE:
-sudo zypper install ./pythinker-code-0.19.0.x86_64.rpm
+sudo zypper install ./pythinker-code-0.21.0.x86_64.rpm
 
 # Fedora / RHEL (aarch64)
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.aarch64.rpm
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.aarch64.rpm.sha256
-sha256sum -c pythinker-code-0.19.0.aarch64.rpm.sha256
-sudo dnf install ./pythinker-code-0.19.0.aarch64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.aarch64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.aarch64.rpm.sha256
+sha256sum -c pythinker-code-0.21.0.aarch64.rpm.sha256
+sudo dnf install ./pythinker-code-0.21.0.aarch64.rpm
 ```
 
 Both packages drop a small `/usr/bin/pythinker` launcher that execs the real
@@ -333,8 +330,8 @@ binary under `/usr/lib/pythinker/`, so your `$PATH` stays tidy.
 **Verify before install:**
 
 ```sh
-sha256sum -c pythinker-code_0.19.0_amd64.deb.sha256        # Debian/Ubuntu
-sha256sum -c pythinker-code-0.19.0.x86_64.rpm.sha256       # Fedora/RHEL
+sha256sum -c pythinker-code_0.21.0_amd64.deb.sha256        # Debian/Ubuntu
+sha256sum -c pythinker-code-0.21.0.x86_64.rpm.sha256       # Fedora/RHEL
 ```
 
 **Upgrade:** download the new `.deb`/`.rpm` from Releases and `dpkg -i` /
@@ -364,7 +361,7 @@ at `~/.local/bin/pythinker`.
 curl -fsSL https://pythinker.com/install.sh | bash
 
 # Pin a specific version
-curl -fsSL https://pythinker.com/install.sh | bash -s -- --version 0.19.0
+curl -fsSL https://pythinker.com/install.sh | bash -s -- --version 0.21.0
 
 # Custom prefix (defaults to $HOME/.local)
 curl -fsSL https://pythinker.com/install.sh | bash -s -- --prefix /opt/pythinker

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/README.md

@@ -14,7 +14,7 @@
 [![License: Apache-2.0](https://img.shields.io/badge/License-Apache--2.0-16a34a.svg?style=for-the-badge)](https://github.com/mohamed-elkholy95/Pythinker-Code/blob/main/LICENSE)
 [![CI](https://img.shields.io/github/actions/workflow/status/mohamed-elkholy95/Pythinker-Code/ci-pythinker-cli.yml?branch=main&label=CI&style=for-the-badge&logo=githubactions&logoColor=white)](https://github.com/mohamed-elkholy95/Pythinker-Code/actions/workflows/ci-pythinker-cli.yml?query=branch%3Amain)
 
-[![Downloads](https://img.shields.io/pepy/dt/pythinker-code?cacheSeconds=300)](https://pepy.tech/projects/pythinker-code)
+[![PyPI downloads](https://assets.piptrends.com/get-last-month-downloads-badge/pythinker-code.svg)](https://piptrends.com/package/pythinker-code)
 [![Code style: Ruff](https://img.shields.io/badge/code%20style-ruff-f59e0b.svg?style=flat-square&logo=ruff&logoColor=white)](https://docs.astral.sh/ruff/)
 [![ACP ready](https://img.shields.io/badge/ACP-ready-7c3aed.svg?style=flat-square)](https://github.com/agentclientprotocol/agent-client-protocol)
 [![MCP tools](https://img.shields.io/badge/MCP-tools-0891b2.svg?style=flat-square)](https://modelcontextprotocol.io/)
@@ -50,14 +50,11 @@ It speaks the [**Agent Client Protocol (ACP)**](https://github.com/agentclientpr
 
 ---
 
-## 🆕 What's New in 0.19.0
+## 🆕 What's New in 0.21.0
 
-- **Central per-project agent memory.** Pythinker now keeps a durable per-project memory (`MEMORY.md` / `USER.md`) under `~/.pythinker/projects/<key>/memory/`, written through a new root-only `Memory` tool with content guards and secret-shape detection, recalled into the root agent's first wakeup prompt within a bounded budget, and inspectable with the new `/memory` command.
-- **Non-blocking update flow.** The blocking pre-start update prompt is replaced by a cached, no-network startup notice plus a triggerable `/update` command, and the native Windows installer now waits on the launching process before swapping files and cleans up its staged installer.
-- **Fully native Homebrew formula.** Brew installs are generated from the same GitHub Release tarballs as the curl installer (including macOS Intel native assets), so they no longer depend on PyPI virtualenv resources.
-- **Release pipeline hardening.** TestPyPI publishing is now non-blocking and publish steps carry timeouts, so a transient staging flake no longer fails a release.
+- **Non-blocking memory recall.** Recall I/O is now dispatched to a thread-pool executor, preventing the event loop from stalling during memory lookups. Sessions with large memory stores or slow disks will remain responsive throughout recall.
 
-Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.19.0`, or use the native installer for your platform from the [Releases page](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest).
+Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.21.0`, or use the native installer for your platform from the [Releases page](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest).
 
 
 ---
@@ -147,7 +144,7 @@ matches your OS — no Python, Node, or `uv` prerequisite.
 
 | Platform | Recommended install | Artifact source |
 |---|---|---|
-| **🪟 Windows** | `irm https://pythinker.com/install.ps1 \| iex` | `PythinkerSetup-0.19.0.exe` from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
+| **🪟 Windows** | `irm https://pythinker.com/install.ps1 \| iex` | `PythinkerSetup-0.21.0.exe` from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
 | **<img src="https://img.shields.io/badge/-macOS-000000?style=flat-square&logo=apple&logoColor=white" alt="macOS"> / <img src="https://img.shields.io/badge/-Linux-FCC624?style=flat-square&logo=linux&logoColor=black" alt="Linux">** | `curl -fsSL https://pythinker.com/install.sh \| bash` | native tarball from [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
 | **<img src="https://img.shields.io/badge/-macOS-000000?style=flat-square&logo=apple&logoColor=white" alt="macOS"> — Homebrew** | `brew install mohamed-elkholy95/pythinker/pythinker-code` | auto-published Homebrew tap |
 | **<img src="https://img.shields.io/badge/-Linux-FCC624?style=flat-square&logo=linux&logoColor=black" alt="Linux"> — system package** | Download the `.deb` or `.rpm` for your distro below | [Releases](https://github.com/mohamed-elkholy95/Pythinker-Code/releases/latest) |
@@ -172,7 +169,7 @@ pythinker                      # start the interactive TUI
 
 ### 🪟 Windows — native installer
 
-`PythinkerSetup-0.19.0.exe` is a signed* Inno Setup wizard. Installs per-user
+`PythinkerSetup-0.21.0.exe` is a signed* Inno Setup wizard. Installs per-user
 into `%LOCALAPPDATA%\Programs\Pythinker`, registers `pythinker` on your user
 PATH (`HKCU\Environment`), broadcasts `WM_SETTINGCHANGE` so new shells see
 the change. **No UAC prompt.**
@@ -183,13 +180,13 @@ irm https://pythinker.com/install.ps1 | iex
 
 # Or manually download the installer + checksum from the Releases page,
 # verify with Get-FileHash, then run:
-.\PythinkerSetup-0.19.0.exe
+.\PythinkerSetup-0.21.0.exe
 
 # Open a fresh PowerShell
 pythinker --version
 ```
 
-**Per-machine install** (IT-managed boxes): `.\PythinkerSetup-0.19.0.exe /ALLUSERS`
+**Per-machine install** (IT-managed boxes): `.\PythinkerSetup-0.21.0.exe /ALLUSERS`
 installs to `%ProgramFiles%\Pythinker` and writes PATH to HKLM (requires admin).
 
 **Upgrade:** `pythinker update` from inside the running app — it downloads
@@ -240,26 +237,26 @@ attached to every GitHub Release.
 
 ```sh
 # Debian / Ubuntu (x86_64)
-sudo dpkg -i pythinker-code_0.19.0_amd64.deb
+sudo dpkg -i pythinker-code_0.21.0_amd64.deb
 sudo apt-get install -f       # only if dpkg reports missing deps
 
 # Debian / Ubuntu (ARM64)
-sudo dpkg -i pythinker-code_0.19.0_arm64.deb
+sudo dpkg -i pythinker-code_0.21.0_arm64.deb
 
 # Fedora / RHEL / openSUSE (x86_64)
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.x86_64.rpm
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.x86_64.rpm.sha256
-sha256sum -c pythinker-code-0.19.0.x86_64.rpm.sha256
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.x86_64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.x86_64.rpm.sha256
+sha256sum -c pythinker-code-0.21.0.x86_64.rpm.sha256
 # Fedora / RHEL:
-sudo dnf install ./pythinker-code-0.19.0.x86_64.rpm
+sudo dnf install ./pythinker-code-0.21.0.x86_64.rpm
 # openSUSE:
-sudo zypper install ./pythinker-code-0.19.0.x86_64.rpm
+sudo zypper install ./pythinker-code-0.21.0.x86_64.rpm
 
 # Fedora / RHEL (aarch64)
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.aarch64.rpm
-curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.19.0/pythinker-code-0.19.0.aarch64.rpm.sha256
-sha256sum -c pythinker-code-0.19.0.aarch64.rpm.sha256
-sudo dnf install ./pythinker-code-0.19.0.aarch64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.aarch64.rpm
+curl -LO https://github.com/mohamed-elkholy95/Pythinker-Code/releases/download/v0.21.0/pythinker-code-0.21.0.aarch64.rpm.sha256
+sha256sum -c pythinker-code-0.21.0.aarch64.rpm.sha256
+sudo dnf install ./pythinker-code-0.21.0.aarch64.rpm
 ```
 
 Both packages drop a small `/usr/bin/pythinker` launcher that execs the real
@@ -268,8 +265,8 @@ binary under `/usr/lib/pythinker/`, so your `$PATH` stays tidy.
 **Verify before install:**
 
 ```sh
-sha256sum -c pythinker-code_0.19.0_amd64.deb.sha256        # Debian/Ubuntu
-sha256sum -c pythinker-code-0.19.0.x86_64.rpm.sha256       # Fedora/RHEL
+sha256sum -c pythinker-code_0.21.0_amd64.deb.sha256        # Debian/Ubuntu
+sha256sum -c pythinker-code-0.21.0.x86_64.rpm.sha256       # Fedora/RHEL
 ```
 
 **Upgrade:** download the new `.deb`/`.rpm` from Releases and `dpkg -i` /
@@ -299,7 +296,7 @@ at `~/.local/bin/pythinker`.
 curl -fsSL https://pythinker.com/install.sh | bash
 
 # Pin a specific version
-curl -fsSL https://pythinker.com/install.sh | bash -s -- --version 0.19.0
+curl -fsSL https://pythinker.com/install.sh | bash -s -- --version 0.21.0
 
 # Custom prefix (defaults to $HOME/.local)
 curl -fsSL https://pythinker.com/install.sh | bash -s -- --prefix /opt/pythinker

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pythinker-code"
-version = "0.19.0"
+version = "0.21.0"
 description = "Pythinker Code is your next CLI agent."
 readme = "README.md"
 requires-python = ">=3.12"

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/CHANGELOG.md

@@ -15,6 +15,26 @@ GitHub Releases page; `0.8.0` is the new starting line.
 
 ## Unreleased
 
+## 0.21.0 (2026-05-28)
+
+### What changed in this release
+
+- **Non-blocking memory recall.** Recall I/O is now dispatched to a thread-pool executor, preventing the event loop from stalling during memory lookups. Sessions with large memory stores or slow disks will remain responsive throughout recall.
+
+Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.21.0`, or use the native installer for your OS (see the README install table).
+
+## 0.20.0 (2026-05-28)
+
+### What changed in this release
+
+- **Ranked recall injection and Scratchpad working memory.** The memory subsystem gains a dependency-free lexical retriever, ranked recall injection replacing the previous verbatim memory dump, and a new root-only `Scratchpad` tool that lets the agent record structured notes across tool calls during a session.
+- **Full memory lifecycle: Phases B–D.** A unified injection bus re-arms recall on every Memory or Scratchpad write (Phase B), session-end episodic recaps are harvested before compaction into a per-project journal (Phase C, off by default), and an approval-gated inbox consolidation flow (`/memory inbox`) proposes—but never auto-applies—memory merges (Phase D).
+- **Context-aware and web-fresh agent subagents.** Code-reviewer, security-reviewer, coder, and plan specialists now mandate a context7-first / web-fallback freshness check before scoring third-party findings or writing against a library. The system prompt gains a Granular Todo discipline (one in-progress sub-todo per dispatched subagent) and an Engineering Discipline section.
+- **Memory robustness hardening.** Compaction harvest failures are now isolated per-phase (a broken prepare step no longer silently cancels recall), inbox JSON-parse errors log and skip instead of dropping silently, and the recall injection bus uses a knapsack budget that continues filling with lower-priority candidates after a high-priority slot overflows.
+- **Animated robot logo in all native installers.** The curl-bash and PowerShell native installers now display a Tetris-style animated robot-head logo on supported terminals, with a static fallback for CI, `NO_COLOR`, dumb-term, or `PYTHINKER_NO_ANIMATION=1`.
+
+Upgrade with `pythinker update`, `pip install --upgrade pythinker-code==0.20.0`, or use the native installer for your OS (see the README install table).
+
 ## 0.19.0 (2026-05-27)
 
 ### What changed in this release

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/agent.yaml

@@ -13,6 +13,7 @@ agent:
     - "pythinker_code.tools.ask_user:AskUserQuestion"
     - "pythinker_code.tools.todo:SetTodoList"
     - "pythinker_code.tools.memory:Memory"
+    - "pythinker_code.tools.scratchpad:Scratchpad"
     - "pythinker_code.tools.shell:Shell"
     - "pythinker_code.tools.background:TaskList"
     - "pythinker_code.tools.background:TaskOutput"

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/code_reviewer.yaml

@@ -23,6 +23,12 @@ agent:
       - Build a review context packet: base ref/diff scope, changed behavior, likely tests, user-visible impact, and validation evidence.
       - Flag only issues introduced or made reachable by the diff.
       - Prefer no finding over vague speculation. Every finding must cite concrete evidence and a failure mode.
+
+      Freshness check (run BEFORE flagging third-party library or framework misuse):
+      - For every third-party API, SDK call, framework primitive, or "best practice" the diff turns on, verify the current canonical usage. Prefer a context7 MCP query (e.g. `mcp__context7__query-docs` with the library id) when registered with the parent runtime; otherwise use `SearchWeb` to locate the official documentation and `FetchURL` to read the current page.
+      - Do NOT flag "deprecated", "removed", "wrong API", or "missing parameter" purely from training-cutoff memory. Either verify against the live docs and cite the URL in EVIDENCE, or downgrade the finding to RISKS with a "needs verification" note.
+      - The freshness check is itself read-only and bounded — one or two fetches per load-bearing finding is enough; do not crawl.
+      - Skip the check for purely internal-codebase findings (logic, scope, project conventions); it applies only to third-party-surface claims.
       - Do not request tests unless they cover a distinct behavior or risk introduced by the change.
       - Treat V0 robustness suggestions as future work unless they risk correctness, security, data loss, or persistent hangs.
       - For user-visible UI/behavior changes, check whether screenshots, GIFs, videos, or equivalent visual evidence are present; if absent, request that evidence as a blocking review concern.
@@ -48,6 +54,8 @@ agent:
     - "pythinker_code.tools.file:ReadFile"
     - "pythinker_code.tools.file:Grep"
     - "pythinker_code.tools.skill:ReadSkill"
+    - "pythinker_code.tools.web:SearchWeb"
+    - "pythinker_code.tools.web:FetchURL"
   exclude_tools:
     - "pythinker_code.tools.file:WriteFile"
     - "pythinker_code.tools.file:StrReplaceFile"

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/coder.yaml

@@ -14,6 +14,7 @@ agent:
 
       Implementation method:
       - Before editing, read the target files and confirm the line ranges/patterns you will change.
+      - Before writing against a third-party library, SDK, cloud service, or framework, pull its current API docs first. Prefer a context7 MCP query (e.g. `mcp__context7__query-docs` with the library id) when registered with the parent runtime; otherwise use `SearchWeb` to find the official docs and `FetchURL` to read the current page. Do NOT write API calls from training-cutoff memory for surfaces that move (LLM SDKs, cloud SDKs, web frameworks, ORM/migration tools, anything < 2 years old). Cite the doc URL or context7 result in EVIDENCE.
       - Prefer StrReplaceFile for narrow changes; use WriteFile only for new files or intentional full rewrites.
       - Add or update tests when the brief requires behavior changes and the project has relevant tests.
       - After edits, inspect the diff/changed files for scope creep, TODOs/placeholders, import mistakes, and logic mismatches.

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/plan.yaml

@@ -21,6 +21,12 @@ agent:
       - For every task, include artifacts to change, acceptance criteria, suggested specialist (`explore`, `implementer`, `review`, `security-reviewer`, `debugger`, `verifier`), and the smallest verification command/check.
       - Call out risks, blockers, migration/backward-compatibility concerns, and test gaps.
 
+      Library/API freshness (run BEFORE recommending an external dependency or API surface):
+      - For every third-party library, SDK, framework, or cloud service the plan turns on (new dep, version bump, non-trivial API surface, security-sensitive primitive), pull the current docs first. Prefer a context7 MCP query (e.g. `mcp__context7__query-docs` with the library id) when registered with the parent runtime; otherwise use `SearchWeb` to find the official docs and `FetchURL` to read the current page.
+      - Do NOT plan around an API from training-cutoff memory if it has moved (LLM SDKs, cloud SDKs, web frameworks, ORM/migration tools). Verify the call shape, supported versions, and any documented migration path.
+      - Cite the doc reference inline next to the task that depends on it, in EVIDENCE.
+      - When the freshness check changes the plan (e.g. an API was removed, a new auth flow is mandated), call it out in RISKS as a constraint the implementer must honor.
+
       Ground the plan in evidence. Read enough files to avoid guessing, name the trade-offs, and choose one path with a reason. Each step should name the artifact it changes and the verification that proves it worked. Order steps by dependency first, then by risk reduced per effort.
 
       Final response contract:

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/security_reviewer.yaml

@@ -22,6 +22,13 @@ agent:
       - Check graceful degradation, observability/logging, recovery behavior, structured result/status correctness, and approval/policy mismatches when they affect security posture.
       - Translate JSON output into the structured response block below.
 
+      Latest advisory pull (run BEFORE finalizing severity):
+      - Identify every third-party surface in the diff: dependencies (pyproject/requirements/lock), SDK calls, framework primitives, crypto/auth helpers, network/serialization libs.
+      - For each surface, pull current advisories and release notes. Prefer a context7 MCP query (e.g. `mcp__context7__query-docs` with the library id) when registered with the parent runtime; otherwise use `SearchWeb` for the latest CVE/GHSA/security-advisory bulletin and `FetchURL` for the canonical advisory text.
+      - Never cite a CVE, GHSA id, or "X is patched in vY" claim from memory alone — verify against the live advisory body. If the network is unavailable, omit the citation and record it under RISKS as a coverage gap.
+      - For framework-specific threat patterns, the reference is `blackbox/pythinker-security-scanner` (especially `docs/supported-tech.md` threat highlights and `packages/scanner/src/matchers/`). Cross-check the diff against the relevant tech tag's highlights.
+      - Web fetches must be evidence, not chatter: cite the URL inline in EVIDENCE when a finding turns on a current advisory.
+
       Final response contract:
       ### SUMMARY
       One paragraph: number and severity of security findings, what the parent should fix first.
@@ -39,6 +46,8 @@ agent:
     - "pythinker_code.tools.shell:Shell"
     - "pythinker_code.tools.file:ReadFile"
     - "pythinker_code.tools.file:Grep"
+    - "pythinker_code.tools.web:SearchWeb"
+    - "pythinker_code.tools.web:FetchURL"
   exclude_tools:
     - "pythinker_code.tools.file:WriteFile"
     - "pythinker_code.tools.file:StrReplaceFile"

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/agents/default/system.md

@@ -39,6 +39,37 @@ For any codebase, architecture, debugging, security, performance, planning, or "
 
 **Professional handoff format:** For substantial tasks, keep a visible plan/todo and structure work as `context -> assessment -> plan -> execution -> verification -> residual risks`. Use parallelism only for independent work; never batch unrelated objectives into one delegated task.
 
+# Engineering Discipline
+
+These principles govern every engineering response. They override speed: a slow right answer beats a fast wrong one.
+
+**1. Think before coding — don't assume, don't hide confusion, surface tradeoffs.**
+- State your assumptions explicitly before implementing. If uncertain, ask.
+- If the request admits multiple interpretations, present them — don't pick one silently.
+- If a simpler approach exists than what the user proposed, say so before building the complex one. Push back when warranted.
+- If something is unclear, stop. Name what's confusing. Ask one focused question.
+- Clarifying questions belong **before** implementation, not after mistakes.
+
+**2. Simplicity first — minimum code that solves the problem, nothing speculative.**
+- No features beyond what was asked.
+- No abstractions for single-use code.
+- No "flexibility" or "configurability" that wasn't requested.
+- No error handling for impossible scenarios; validate at boundaries only.
+- If a 200-line draft could be 50 lines, rewrite it before showing it.
+- Self-check: *"Would a senior engineer call this overcomplicated?"* If yes, simplify.
+
+**3. Goal-driven execution — define success criteria, then loop until verified.**
+- Transform vague tasks into verifiable goals before writing code:
+  - "Add validation" → "Write tests for invalid inputs, then make them pass."
+  - "Fix the bug" → "Write a test that reproduces it, then make it pass."
+  - "Refactor X" → "Ensure tests pass before and after; behavior identical."
+  - "Make it faster" → "Benchmark current, set target, prove improvement on same inputs."
+- For multi-step work, state the plan inline as `Step → verify: check`, then execute against it.
+- "It compiles" is not verification. "It type-checks" is not verification. Verification is a passing test, a working repro, or a deterministic command that confirms the intended behavior.
+- Don't claim done without proof. If verification can't run, say so explicitly under BLOCKERS instead of asserting success.
+
+These principles are working if: diffs contain only requested changes, fewer rewrites land because of overcomplication, and clarifying questions appear before the first edit rather than after the first mistake.
+
 # Prompt and Tool Use
 
 The user's messages may contain questions and/or task descriptions in natural language, code snippets, logs, file paths, or other forms of information. Read them, understand them and do what the user requested. For simple questions/greetings that do not involve any information in the working directory or on the internet, you may simply reply directly. For anything else, default to taking action with tools. When the request could be interpreted as either a question to answer or a task to complete, treat it as a task.
@@ -59,6 +90,8 @@ For any non-trivial request, decompose before acting:
 
 - Preview the terrain first: scan the directory structure, file headers, and relevant module boundaries before choosing an implementation path.
 - Use `SetTodoList` for multi-step work so the user can see the active plan and progress.
+- **Granular todos, not umbrella todos.** Each todo must name a single concrete deliverable a human can recognize as "this part is done." Avoid umbrella titles like "Determine X" or "Investigate Y" that cover hours of parallel work — they freeze the progress UI while real work happens underneath. If a single todo would stay `in_progress` for more than ~3 minutes, it is too coarse: split it before launching work.
+- **One todo per dispatched child.** When you launch `RunAgents` with N children, the visible todo list MUST contain one in_progress sub-todo per child (or per independent objective the batch covers) **before** the batch starts. Update each sub-todo to `done` as that child returns — do not wait for the whole batch to finish to flip a single umbrella todo. Same rule applies to multiple parallel `Agent` calls in the same turn.
 - Split broad work into independent chunks; use parallel tool calls or focused subagents for chunks that do not depend on each other.
 - For large codebase scans, start with indexes/graphs and targeted searches; avoid one vague repo-wide subagent prompt. If using background agents for thorough exploration, set a realistic explicit timeout and keep scopes narrow. If agents time out, do not repeat the same broad launch; summarize partial evidence, run targeted direct scans, and resume or relaunch narrower agents only when useful.
 - Re-read the plan after each phase and adjust it when new evidence changes the approach.

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/app.py

@@ -345,13 +345,11 @@ class PythinkerCLI:
 
         soul = PythinkerSoul(agent, context=context)
 
-        from pythinker_code.project_memory import (
-            ProjectMemoryInjectionProvider,
-            ProjectMemoryStore,
-        )
+        from pythinker_code.memory.recall import RecallInjectionProvider
+        from pythinker_code.project_memory import ProjectMemoryStore
 
         soul.add_injection_provider(
-            ProjectMemoryInjectionProvider(ProjectMemoryStore(session.work_dir))
+            RecallInjectionProvider(ProjectMemoryStore(session.work_dir), session)
         )
 
         # Activate plan mode if requested (for new sessions or --plan flag)

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/cli/__init__.py

@@ -907,7 +907,20 @@ def pythinker(
         # Session scratchpads are retained as compact history for future recall.
         # ``cleanup_scratchpad`` is kept for call-site compatibility but no longer
         # triggers automatic deletion after a successful run.
-        _ = cleanup_scratchpad, exit_code
+        _ = cleanup_scratchpad
+        if exit_code == ExitCode.SUCCESS and getattr(
+            getattr(config, "memory", None), "journal_recaps", False
+        ):
+            with contextlib.suppress(Exception):
+                from pythinker_code.memory.recap import build_session_recap
+                from pythinker_code.project_memory import ProjectMemoryStore
+
+                recap = build_session_recap(
+                    state=last_session.state,
+                    session_id=last_session.id,
+                    request=last_session.title,
+                )
+                await ProjectMemoryStore(last_session.work_dir).append_journal(recap)
         _print_resume_hint(last_session)
         if last_session.is_empty():
             # Always clean up empty sessions regardless of exit code

{pythinker_code-0.19.0 → pythinker_code-0.21.0}/src/pythinker_code/config.py

@@ -135,6 +135,38 @@ class NotificationConfig(BaseModel):
     claim_stale_after_ms: int = Field(default=15_000, ge=1000)
 
 
+class MemoryConfig(BaseModel):
+    """Context- and memory-aware agent configuration."""
+
+    lexical_recall: bool = Field(
+        default=True,
+        description="Use relevance-ranked lexical recall instead of verbatim project memory.",
+    )
+    injection_bus: bool = Field(
+        default=True,
+        description="Budget dynamic prompt injections through the injection bus.",
+    )
+    injection_ceiling_tokens: int = Field(
+        default=2048,
+        ge=256,
+        description="Maximum total tokens reserved for dynamic prompt injections.",
+    )
+    harvest_on_compaction: bool = Field(
+        default=False,
+        description=(
+            "Persist safe decisions/blockers/next steps before compaction discards history."
+        ),
+    )
+    journal_recaps: bool = Field(
+        default=False,
+        description="Write stable-schema JOURNAL.md session recap blocks.",
+    )
+    consolidation: bool = Field(
+        default=False,
+        description="Enable approval-gated memory inbox consolidation helpers.",
+    )
+
+
 class PythinkerAISearchConfig(BaseModel):
     """Pythinker AI Search configuration."""
 
@@ -324,6 +356,7 @@ class Config(BaseModel):
         default_factory=NotificationConfig, description="Notification configuration"
     )
     services: Services = Field(default_factory=Services, description="Services configuration")
+    memory: MemoryConfig = Field(default_factory=MemoryConfig, description="Memory configuration")
     feedback: FeedbackConfig = Field(
         default_factory=FeedbackConfig,
         description="User-submitted feedback endpoint configuration",

pythinker_code-0.21.0/src/pythinker_code/memory/__init__.py

@@ -0,0 +1 @@
+"""Context- & memory-aware agent subsystem (retrieval, recall, sanitization)."""

pythinker_code-0.21.0/src/pythinker_code/memory/consolidation.py

@@ -0,0 +1,115 @@
+from __future__ import annotations
+
+import json
+import os
+import re
+from dataclasses import asdict, dataclass
+from pathlib import Path
+
+from loguru import logger
+from pythinker_host.path import HostPath
+
+from pythinker_code.memory.recall import gather_candidates
+from pythinker_code.memory.recap import content_hash
+from pythinker_code.project_memory import ProjectMemoryStore
+
+
+@dataclass(frozen=True, slots=True)
+class InboxCandidate:
+    id: str
+    target: str
+    title: str
+    content: str
+    source_path: str
+    content_hash: str
+
+
+def _safe_id(value: str) -> str:
+    return re.sub(r"[^a-z0-9_-]", "-", value.lower())[:32].strip("-") or "candidate"
+
+
+async def inbox_dir(store: ProjectMemoryStore) -> Path:
+    root = await store._ensure_dir()  # pyright: ignore[reportPrivateUsage]
+    path = root / "memory" / "inbox"
+    path.mkdir(parents=True, exist_ok=True)
+    return path
+
+
+async def list_inbox_candidates(store: ProjectMemoryStore) -> list[InboxCandidate]:
+    directory = await inbox_dir(store)
+    out: list[InboxCandidate] = []
+    for path in sorted(directory.glob("*.json")):
+        try:
+            data = json.loads(path.read_text(encoding="utf-8"))
+            out.append(InboxCandidate(**data))
+        except Exception as exc:
+            logger.debug("inbox candidate {} dropped during listing: {!r}", path.name, exc)
+            continue
+    return out
+
+
+async def generate_inbox_candidates(
+    store: ProjectMemoryStore, work_dir: HostPath, *, limit: int = 20
+) -> list[InboxCandidate]:
+    """Stage scratch/journal candidates for approval-gated durable memory consolidation."""
+    existing_entries = [*await store.read_entries("memory"), *await store.read_entries("user")]
+    existing_hashes = {
+        content_hash(tier="memory", title=entry[:60], body=entry) for entry in existing_entries
+    }
+    staged = {candidate.content_hash for candidate in await list_inbox_candidates(store)}
+    directory = await inbox_dir(store)
+    candidates: list[InboxCandidate] = []
+    for block in await gather_candidates(store, work_dir):
+        if block.tier in {"memory", "user"}:
+            continue
+        digest = content_hash(tier="memory", title=block.title, body=block.content)
+        if digest in existing_hashes or digest in staged:
+            continue
+        candidate = InboxCandidate(
+            id=_safe_id(digest),
+            target="memory",
+            title=block.title or block.tier,
+            content=block.content,
+            source_path=block.source_path,
+            content_hash=digest,
+        )
+        path = directory / f"{candidate.id}.json"
+        fd = os.open(path, os.O_WRONLY | os.O_CREAT | os.O_EXCL, 0o600)
+        with os.fdopen(fd, "w", encoding="utf-8") as fh:
+            json.dump(asdict(candidate), fh, ensure_ascii=False, indent=2)
+        candidates.append(candidate)
+        staged.add(digest)
+        if len(candidates) >= limit:
+            break
+    return candidates
+
+
+async def approve_inbox_candidate(store: ProjectMemoryStore, candidate_id: str) -> str:
+    directory = await inbox_dir(store)
+    path = directory / f"{_safe_id(candidate_id)}.json"
+    if not path.is_file():
+        return "Candidate not found."
+    try:
+        data = json.loads(path.read_text(encoding="utf-8"))
+        candidate = InboxCandidate(**data)
+    except Exception as exc:
+        logger.warning(
+            "inbox candidate {} rejected on approve due to parse error: {!r}",
+            path.name,
+            exc,
+        )
+        return "Candidate file is corrupt and cannot be approved."
+    result = await store.add("memory", candidate.content)
+    if not result.ok:
+        return result.message
+    path.unlink(missing_ok=True)
+    return "Candidate approved and added to project memory."
+
+
+async def reject_inbox_candidate(store: ProjectMemoryStore, candidate_id: str) -> str:
+    directory = await inbox_dir(store)
+    path = directory / f"{_safe_id(candidate_id)}.json"
+    if not path.is_file():
+        return "Candidate not found."
+    path.unlink(missing_ok=True)
+    return "Candidate rejected."