chore: promote next to main

.github/workflows/codeql.yml

@@ -31,15 +31,15 @@ jobs:
         uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
 
       - name: Initialize CodeQL
-        uses: github/codeql-action/init@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
+        uses: github/codeql-action/init@db488ddef3bf6cb639b32c2e9a7c0a7ea8271d28 # v4.37.8
         with:
           languages: ${{ matrix.language }}
           queries: security-and-quality
 
       - name: Autobuild
-        uses: github/codeql-action/autobuild@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
+        uses: github/codeql-action/autobuild@db488ddef3bf6cb639b32c2e9a7c0a7ea8271d28 # v4.37.8
 
       - name: Perform CodeQL Analysis
-        uses: github/codeql-action/analyze@5595ccaf912efad79be6eef63a5619ff05969be3 # v4.37.6
+        uses: github/codeql-action/analyze@db488ddef3bf6cb639b32c2e9a7c0a7ea8271d28 # v4.37.8
         with:
           category: "/language:${{ matrix.language }}"

AGENTS.md

@@ -38,7 +38,8 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 
 ### Project memory
 
-<aidd_project_memory>
+<!-- aidd_project_memory:start -->
+
 @aidd_docs/memory/architecture.md
 @aidd_docs/memory/browsing.md
 @aidd_docs/memory/codebase-map.md
@@ -47,7 +48,8 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 @aidd_docs/memory/project-brief.md
 @aidd_docs/memory/testing.md
 @aidd_docs/memory/vcs.md
-</aidd_project_memory>
+
+<!-- aidd_project_memory:end -->
 
 - If the block above is empty, run `ls -1tr aidd_docs/memory/` and read each file.
 - Load `aidd_docs/memory/external/*` when the user asks.

CLAUDE.md

@@ -38,7 +38,8 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 
 ### Project memory
 
-<aidd_project_memory>
+<!-- aidd_project_memory:start -->
+
 @aidd_docs/memory/architecture.md
 @aidd_docs/memory/browsing.md
 @aidd_docs/memory/codebase-map.md
@@ -47,7 +48,8 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 @aidd_docs/memory/project-brief.md
 @aidd_docs/memory/testing.md
 @aidd_docs/memory/vcs.md
-</aidd_project_memory>
+
+<!-- aidd_project_memory:end -->
 
 - If the block above is empty, run `ls -1tr aidd_docs/memory/` and read each file.
 - Load `aidd_docs/memory/external/*` when the user asks.

README.md

@@ -2,22 +2,24 @@
 
 <img src="docs/assets/logo.png" alt="AIDD" width="100" />
 
-# The AI-Driven Dev Framework 🇫🇷
+# AI-Driven Dev Framework
 
-`Open source` agnostic framework **to generate high quality clean code**.
+## Enterprise-grade SDLC for high-quality software development.
 
-_(Already tested on `Legacy` codebases)_
+Unify **engineering teams** around **standardized workflows** and **shared best practices**, across modern stacks and **legacy systems**, while reducing **technical debt**.
 
-[![Made in France](https://img.shields.io/badge/made%20in-France-0055A4?labelColor=EF4135)](https://www.ai-driven-dev.fr/)
+🧱 **IDE agnostic** · 🏗️ **Legacy systems** · 🌱 **Token-optimized** · 🇫🇷 **Made in France**
 
 <p>
-  <!--counts:start--><kbd>7 plugins</kbd> · <kbd>47 skills</kbd> · <kbd>2 agents</kbd><!--counts:end--> · <kbd>MIT</kbd>
+  <!--counts:start--><kbd>7 plugins</kbd> · <kbd>47 skills</kbd> · <kbd>2 agents</kbd><!--counts:end-->
 </p>
 
-[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![Open Source](https://img.shields.io/badge/Open_Source-Yes-yellow?logo=open-source-initiative&logoColor=white)](https://opensource.org/)
+[![License: MIT](https://img.shields.io/badge/License-MIT-purple.svg)](LICENSE)
 [![Latest Release](https://img.shields.io/github/v/release/ai-driven-dev/framework?include_prereleases&sort=semver)](https://github.com/ai-driven-dev/framework/releases)
 [![CI](https://github.com/ai-driven-dev/framework/actions/workflows/ci.yml/badge.svg?branch=main)](https://github.com/ai-driven-dev/framework/actions/workflows/ci.yml)
 
+
 <p>🗺️ <a href="https://github.com/orgs/ai-driven-dev/projects/8"><b>Live roadmap</b></a></p>
 
 </div>
@@ -195,18 +197,36 @@ flowchart TD
     end
 
     subgraph loop["② Per feature — repeat"]
-      direction LR
-      B["brainstorm"] --> P["plan"] --> I["implement"] --> R["review"] --> C["commit"] --> PR(["✅ PR"])
+      direction TB
+      subgraph frame["Frame"]
+        direction LR
+        N["functional need<br/>or User Story"] --> T["issue / ticket"] --> P["plan"]
+      end
+      subgraph deliver["Deliver"]
+        direction LR
+        I["implement"] --> V["validate"] --> R["review"]
+      end
+      subgraph ship["Ship"]
+        direction LR
+        PR["PR"] --> S(["✅ production"]) --> L["learn"]
+      end
+      P --> I
+      R --> PR
     end
 
-    Onboard --> Memory --> B
+    Onboard --> Memory --> N
+    L -. next need .-> N
 
     classDef hub fill:#D97757,stroke:#9c4f37,color:#fff;
     classDef done fill:#2ea043,stroke:#1a7f37,color:#fff;
     class Onboard hub;
-    class PR done;
+    class S done;
 ```
 
+Start with a functional need or User Story, then track it as an issue or ticket
+before planning. Shipping follows the project's own delivery process. Capture a
+learning only when it is durable enough to improve the next feature.
+
 > 🍳 **More flows** → bundled recipes: [start a project](plugins/aidd-context/skills/12-cook/assets/recipes/start-a-project.md), [ship a feature](plugins/aidd-context/skills/12-cook/assets/recipes/ship-a-feature.md), and more.
 
 ## 🧩 Plugins

aidd_docs/CONTRIBUTING.md

@@ -26,7 +26,7 @@ The skill contract itself lives in one file: `aidd-context:04-skill-generate/ref
 
 ## Syncing Across Tools
 
-If the project uses multiple AI tools (e.g. Claude Code plus Cursor), the same content must be available to each. The memory bank is shared automatically via the `<aidd_project_memory>` block kept in sync by `aidd-context:02-project-memory`. Skills are loaded per-plugin by the runtime, so any skill installed via the marketplace is available across tools that support skills.
+If the project uses multiple AI tools (e.g. Claude Code plus Cursor), the same content must be available to each. The memory bank is shared automatically via the project memory block kept in sync by `aidd-context:02-project-memory`. Skills are loaded per-plugin by the runtime, so any skill installed via the marketplace is available across tools that support skills.
 
 When tools differ in syntax (frontmatter, slash command name, references), follow the IDE mapping reference shipped with each plugin.
 

aidd_docs/README.md

@@ -84,7 +84,7 @@ my-project/
 
 ### Memory Block Lifecycle
 
-Each AI context file (`CLAUDE.md`, `AGENTS.md`, `.github/copilot-instructions.md`, etc.) contains an `<aidd_project_memory>` block. It is:
+Each AI context file (`CLAUDE.md`, `AGENTS.md`, `.github/copilot-instructions.md`, etc.) contains a project memory block. It is:
 
 1. **Seeded** the first time by `aidd-context:02-project-memory` (the skill creates the block if absent).
 2. **Kept in sync** automatically by a session-start hook (`aidd-context/hooks/update_memory.js`) that scans `aidd_docs/memory/` and writes the current list of `.md` files into the block.

aidd_docs/tasks/2026_08/2026_08_28_memory-block-markers/brainstorm.md

@@ -0,0 +1,26 @@
+# Memory block markers that cannot silently stop loading
+
+The `<aidd_project_memory>` block that `update_memory.js` writes into every AI context file is delimited by a bare XML-style tag on its own line. Markdown treats that as an HTML block running until the next blank line, and Claude Code's import parser skips imports inside it exactly as it skips a fenced code block. Every `@aidd_docs/memory/*.md` line the hook generates is therefore ignored, and the project memory never loads. The failure is invisible from inside a session: the agent starts, nothing is reported, and it simply reasons without the memory bank.
+
+The fix is to stop delimiting the block with a construct whose correctness depends on an adjacent blank line, and use HTML comment markers instead. A comment closes on its own line, so nothing following it can ever be captured, whatever a formatter or a future rewrite does to the surrounding whitespace. The same hook already uses that convention for the memory README index. Existing installed repositories carry the old tag, so the change is only complete if the hook migrates them; otherwise the fix ships and those repositories stay dark while the changelog claims otherwise.
+
+## What Is Clear
+
+- Reported in issue #719 and confirmed independently: `CLAUDE.md:41-50` in this repository holds the glued shape, and a session loading it shows the `@` lines unexpanded, with eight memory files absent.
+- Single emitter: `buildBlockContent` in `plugins/aidd-context/hooks/update_memory.js`. `updateMarkers` splices between two constant strings and needs no change when those strings change.
+- The bug is the tag shape, not the content. `buildBlockContent` already pushes a blank line before the on-demand note, so that blank line terminates the HTML block and the on-demand list renders correctly. Only the `@` lines above it are swallowed. That asymmetry is the signature.
+- Chosen shape: the existing `### Project memory` heading for human readers, with `<!-- aidd_project_memory:start -->` and `<!-- aidd_project_memory:end -->` as the machine anchors. The heading is prose the hook does not own; the markers stay the anchor.
+- Rejected: adding a blank line after the opening tag. It is one character and needs no migration, but it is a load-bearing convention nothing in the file declares, and a rewrite already reverted it once.
+- Migration is part of the fix, not a follow-up. The hook returns `null` when the close marker is absent and skips the file with no output, so every already-installed repository would silently stop syncing.
+- The tag name is load-bearing beyond the hook and those references move with it: `plugins/aidd-context/skills/00-onboard/references/state/zones.md`, `plugins/aidd-context/skills/11-explore/actions/01-survey.md`, and the `02-project-memory` templates for `AGENTS.md`, `README.md`, and `memory/README.md`.
+- A regression guard stays in scope: walk a context file line by line, track whether a line opens an HTML block, and fail when an `@` import is found inside one. Comment markers make this specific regression near-impossible, but the guard covers hand-written context files and future generators.
+- Out of scope, its own ticket: generating a `CLAUDE.md` that contains `@AGENTS.md` for repositories that only have `AGENTS.md`. It adds a capability rather than fixing a defect, and it drags in the separate question of which context files the project owns.
+
+## Still Open
+
+- Assumption, unverified: the XML tag currently buys nothing the model uses. Whether the wrapper survives import resolution into the prompt in a form the model reads has not been checked. If it does, the heading in the chosen shape is the replacement cue.
+- Whether the CLI test fixtures under `cli/tests/fixtures/**` that embed the old tag should be migrated with the rest or left pinned to the legacy shape as regression material.
+
+## Next Move
+
+Turn this into a change: new markers in the hook, a one-pass rewrite of the legacy tag before the splice, the template and documentation references moved to the new shape, and a check that fails when an import sits inside an HTML block.

cli/CLAUDE.md

@@ -37,7 +37,8 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 
 ### Project memory
 
-<aidd_project_memory>
+<!-- aidd_project_memory:start -->
+
 @aidd_docs/memory/architecture.md
 @aidd_docs/memory/auth.md
 @aidd_docs/memory/cli.md
@@ -47,7 +48,11 @@ Project docs, memory, specs, and plans live in `aidd_docs/`.
 @aidd_docs/memory/project-brief.md
 @aidd_docs/memory/testing.md
 @aidd_docs/memory/vcs.md
-</aidd_project_memory>
+
+<!-- read on demand, not auto-loaded -->
+- aidd_docs/memory/internal/decisions/self-update-version-source-npm.md
+
+<!-- aidd_project_memory:end -->
 
 - If the block above is empty, run `ls -1tr aidd_docs/memory/` and read each file.
 - Load `aidd_docs/memory/external/*` when the user asks.

cli/aidd_docs/memory/README.md

@@ -4,7 +4,7 @@ Structured context the AI assistant reads at the start of a session, so it does
 
 ## How it loads
 
-- The files at the root of `memory/` are referenced by the `<aidd_project_memory>` block in the AI context file and load every session.
+- The files at the root of `memory/` are referenced by the project memory block in the AI context file and load every session.
 - `internal/` and `external/` are listed there too, but load on demand, only when relevant.
 
 ## Files
@@ -21,6 +21,10 @@ The list below is refreshed automatically by the memory hook. Do not edit it by
 - [project-brief.md](project-brief.md)
 - [testing.md](testing.md)
 - [vcs.md](vcs.md)
+
+Read on demand:
+
+- [internal/decisions/self-update-version-source-npm.md](internal/decisions/self-update-version-source-npm.md)
 <!-- files:end -->
 
 ## How to maintain it