tillerhand 3.54.0 → 3.55.0

package/AGENTS.md

@@ -13,7 +13,7 @@ This file holds the shared rules of the project. Claude Code, Codex and Antigrav
 1. **The user decides.** Ask with AskUserQuestion for product, technical, design, content, data, security and cost decisions. Offer 2-4 options; put your recommendation first and label it "(Recommended)". For each option, state when it is better and what it costs. Ask design decisions with a visual preview. In question mode, pick the recommended option on closed topics and record it as `automatic` in the decision log ("Auto decisions" in the status block). **Read `docs/workflow/decision-protocol.md` before the first decision question.** Earlier decisions are in `docs/decisions/log.md`; do not ask them again. Classify every decision as a one-way or two-way door; for one-way doors (database, auth, payment, data model), get the board's opinion first.
 2. **The expert board reviews** (`docs/workflow/board.md`, rules in `.claude/board/rules.md`). The relevant members meet before plan approval, when a task ends, at every commit (light) and at the end of a phase. The board only advises; if there is a REJECT (critical finding), the work is not done and the user is asked.
 3. **Right source.** Do not use information that can change from memory: versions, prices, quotas, regulation, APIs. Use the `source-research` skill (primary source first, every claim with source and date). When using a library, first fetch the current docs with the **Context7** MCP.
-4. **Spend few tokens.** If the code map (codebase-memory-mcp) is connected, query it first when exploring code (`get_architecture`, `search_graph`, `trace_path`, `get_code_snippet`). Do not read files in bulk. Read docs only when needed (table below). Convert a PDF or Office file with the `document-to-markdown` skill first, then read only the part you need.
+4. **Spend few tokens.** If the code map (codebase-memory-mcp) is connected, query it first when exploring code (`get_architecture`, `search_graph`, `trace_path`, `get_code_snippet`). Do not read files in bulk. Read docs only when needed (table below). Convert a PDF or Office file with the `document-to-markdown` skill first, then read only the part you need. When several reads or searches are needed, run them together in one shell call with an inline script and print only what is needed. This is for reading only: writing, deleting, commits and deploys are not done through it. Keep the script inline, not a script file, because guards only see the command text.
 5. **Follow the structure (mandatory).** Write each new file to its place in the `.claude/structure.json` map (by feature: `features/<feature>/`); it is explained in `docs/architecture/structure.md`. Writing outside the map is asked about; clutter files (`*.bak`, `*-old.*`, copies) are blocked. Temporary and trial work goes to `sandbox/`, outdated docs to `docs/archive/`. Changing the map is a technical decision.
 6. **Plan first, then code.** Work that touches more than 2 files goes through `/plan-task`. One task at a time. Do not widen the scope. Four principles: (1) state your assumption, ask if unsure; (2) simplest solution, no flexibility nobody asked for; (3) touch only the lines the task needs; (4) write a verifiable success criterion when you start.
 7. **Root cause, evidence, tests.** Acceptance tests (criterion id in the test name, `AC-01.1`) and adversarial tests are written before the code and are red first. A separate agent (`test-writer`, with the test role's model) writes the tests, not the one who writes the code; the code writer does not touch test files (test protection blocks it). Risky features ship behind a feature flag (`docs/operations/feature-flags.md`). When fixing a bug, reproduce it first. Show that it works before saying "done". If you cannot fix the same error in 2 attempts, stop and ask for direction.

package/README.md

@@ -6,7 +6,7 @@
 </p>
 
 <p align="center">
-  <a href="kit/CHANGELOG.md"><img alt="Version" src="https://img.shields.io/badge/version-3.54.0-2a78d6?style=flat-square"></a>
+  <a href="kit/CHANGELOG.md"><img alt="Version" src="https://img.shields.io/badge/version-3.55.0-2a78d6?style=flat-square"></a>
   <a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/license-MIT-2a78d6?style=flat-square"></a>
   <img alt="Claude Code" src="https://img.shields.io/badge/Claude%20Code-ready-d97757?style=flat-square">
   <img alt="Also Codex and Antigravity" src="https://img.shields.io/badge/also-Codex%20%C2%B7%20Antigravity-5b6577?style=flat-square">

package/README.tr.md

@@ -6,7 +6,7 @@
 </p>
 
 <p align="center">
-  <a href="kit/CHANGELOG.md"><img alt="Sürüm" src="https://img.shields.io/badge/s%C3%BCr%C3%BCm-3.54.0-2a78d6?style=flat-square"></a>
+  <a href="kit/CHANGELOG.md"><img alt="Sürüm" src="https://img.shields.io/badge/s%C3%BCr%C3%BCm-3.55.0-2a78d6?style=flat-square"></a>
   <a href="LICENSE"><img alt="Lisans: MIT" src="https://img.shields.io/badge/lisans-MIT-2a78d6?style=flat-square"></a>
   <img alt="Claude Code" src="https://img.shields.io/badge/Claude%20Code-haz%C4%B1r-d97757?style=flat-square">
   <img alt="Codex ve Antigravity" src="https://img.shields.io/badge/ayr%C4%B1ca-Codex%20%C2%B7%20Antigravity-5b6577?style=flat-square">

package/kit/CHANGELOG.md

@@ -3,6 +3,12 @@
 The kit version is in `kit/VERSION`. To update a project, run `/kit-update` in it.
 Items with an "Adapt by hand" note concern files owned by the project; the update does not touch them.
 
+## [3.55.0] - 2026-11-01
+
+### Added
+- **Reads are batched.** When several reads or searches are needed, the agent runs them together in one shell call with an inline script and prints only what is needed. Writing, deleting, commits and deploys stay out of it, because guards only see the command text.
+- **The AI add-on guide covers programmatic tool calling** for the products you build: how it works, when not to use it and how to measure it.
+
 ## [3.54.0] - 2026-11-01
 
 ### Added

package/kit/VERSION

@@ -1 +1 @@
-3.54.0
+3.55.0

package/kit/guide.html

@@ -361,7 +361,7 @@ footer .container { display: flex; flex-wrap: wrap; gap: 16px; justify-content:
 <section class="hero">
   <div class="container">
     <div>
-      <span class="pill"><b>v3.54.0</b> Starter kit for Claude Code · MIT</span>
+      <span class="pill"><b>v3.55.0</b> Starter kit for Claude Code · MIT</span>
       <h1>The AI builds.<br><em>You steer.</em></h1>
       <p class="sub">TillerHand takes a product from idea to production and maintenance with Claude Code. At every important decision the AI gives you the options, its recommendation and the cost of each option. An expert board reviews the work, and guardrails stop mistakes before they happen.</p>
       <div class="setup">
@@ -1255,7 +1255,7 @@ Pending decision: email service</pre>
 <footer>
   <div class="container">
     <a class="brand" href="#top" style="font-size:16px"><svg class="mark" style="width:22px;height:22px"><use href="#mark"/></svg>TillerHand</a>
-    <span>Version 3.54.0 · <a href="https://github.com/burakdogrul/TillerHand">GitHub</a> · <a href="https://github.com/burakdogrul/TillerHand/blob/main/LICENSE">MIT license</a> · The KVKK section is a checklist, not legal advice.</span>
+    <span>Version 3.55.0 · <a href="https://github.com/burakdogrul/TillerHand">GitHub</a> · <a href="https://github.com/burakdogrul/TillerHand/blob/main/LICENSE">MIT license</a> · The KVKK section is a checklist, not legal advice.</span>
   </div>
 </footer>
 

package/kit/guide.tr.html

@@ -361,7 +361,7 @@ footer .container { display: flex; flex-wrap: wrap; gap: 16px; justify-content:
 <section class="hero">
   <div class="container">
     <div>
-      <span class="pill"><b>v3.54.0</b> Claude Code için başlangıç kiti · MIT</span>
+      <span class="pill"><b>v3.55.0</b> Claude Code için başlangıç kiti · MIT</span>
       <h1>Kodu AI yazar.<br><em>Dümen sende.</em></h1>
       <p class="sub">TillerHand, Claude Code ile bir ürünü fikirden canlıya ve bakıma kadar götürür. Her önemli kararda AI sana seçenekleri, önerisini ve her seçeneğin bedelini sunar. Bir uzman kurulu işi denetler, korumalar hatayı daha olmadan durdurur.</p>
       <div class="setup">
@@ -1255,7 +1255,7 @@ Bekleyen karar: e-posta servisi</pre>
 <footer>
   <div class="container">
     <a class="brand" href="#top" style="font-size:16px"><svg class="mark" style="width:22px;height:22px"><use href="#mark"/></svg>TillerHand</a>
-    <span>Sürüm 3.54.0 · <a href="https://github.com/burakdogrul/TillerHand">GitHub</a> · <a href="https://github.com/burakdogrul/TillerHand/blob/main/LICENSE">MIT lisansı</a> · KVKK bölümü kontrol listesidir, hukuki görüş değildir.</span>
+    <span>Sürüm 3.55.0 · <a href="https://github.com/burakdogrul/TillerHand">GitHub</a> · <a href="https://github.com/burakdogrul/TillerHand/blob/main/LICENSE">MIT lisansı</a> · KVKK bölümü kontrol listesidir, hukuki görüş değildir.</span>
   </div>
 </footer>
 

package/kit/templates/addons/ai/docs/ai/guide.md

@@ -40,6 +40,17 @@ When a prompt or model changes, "I think it got better" is not enough.
 - [ ] Prompt caching is used for repeated long context (system prompt, documents).
 - [ ] Every call is logged: feature, model, input/output tokens, duration, estimated cost, user id (not the content).
 
+## 4a. Programmatic tool calling
+
+For an AI product you build that calls many tools. It is an API feature, not a Claude Code setting: the model writes code that calls your tools inside a code execution container, so intermediate results stay out of the context and only the final output comes back. Your app still runs each tool.
+
+- **How:** add the `code_execution` tool and set `allowed_callers` on each tool that code may call. Check the current tool version and the supported models in the docs before using it.
+- **When not to use it:** few calls with small results, where container overhead can exceed the saving; a chain where each call depends on reasoning over the last result; a user who needs to see results between calls.
+- **Measure before turning it on broadly:** compare billed input tokens with and without `allowed_callers` on your own workload.
+- `allowed_callers` is not a security boundary: still handle a direct call to any tool.
+
+Source: platform.claude.com/docs/en/agents-and-tools/tool-use/programmatic-tool-calling (2026-11-01).
+
 ## 5. Security
 
 - [ ] The API key is on the server only; the client does not call the provider directly.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "tillerhand",
-  "version": "3.54.0",
+  "version": "3.55.0",
   "description": "Vibe coding kit for Claude Code, Codex and Antigravity: rules, commands, guards and project templates.",
   "license": "MIT",
   "type": "module",