feat(aidd-context): connect services and check the assistant's access

README.md

@@ -11,7 +11,7 @@ Unify **engineering teams** around **standardized workflows** and **shared best
 🧱 **IDE agnostic** · 🏗️ **Legacy systems** · 🌱 **Token-optimized** · 🇫🇷 **Made in France**
 
 <p>
-  <!--counts:start--><kbd>9 plugins</kbd> · <kbd>51 skills</kbd> · <kbd>2 agents</kbd><!--counts:end-->
+  <!--counts:start--><kbd>9 plugins</kbd> · <kbd>52 skills</kbd> · <kbd>2 agents</kbd><!--counts:end-->
 </p>
 
 [![Open Source](https://img.shields.io/badge/Open_Source-Yes-yellow?logo=open-source-initiative&logoColor=white)](https://opensource.org/)
@@ -236,7 +236,7 @@ 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: [MCP installations](plugins/aidd-context/skills/12-cook/assets/recipes/mcp-installation.md), [token optimization](plugins/aidd-context/skills/12-cook/assets/recipes/token-optimization.md), and [installing AIDDy](plugins/aidd-context/skills/12-cook/assets/recipes/install-aiddy-in-codex.md).
+> 🍳 **More flows** → bundled recipes: [Connect your services](plugins/aidd-context/skills/12-cook/assets/recipes/connect-services.md), [token optimization](plugins/aidd-context/skills/12-cook/assets/recipes/token-optimization.md), and [installing AIDDy](plugins/aidd-context/skills/12-cook/assets/recipes/install-aiddy-in-codex.md).
 
 ## 🧩 Plugins
 
@@ -248,7 +248,7 @@ Nine plugins covering the whole SDLC — **install the six stable ones**; they w
 
 ### 🧭 [aidd-context](plugins/aidd-context/README.md)
 
-`13 skills` · stable
+`14 skills` · stable
 
 Project init, memory bank, context-artifact generation, diagrams, learning, exploration.
 
@@ -338,7 +338,7 @@ Full catalog → [`CATALOG.md`](docs/CATALOG.md).
 
 | | |
 | --- | --- |
-| 🍳 **Recipes** | Bundled how-to sheets: [MCP installations](plugins/aidd-context/skills/12-cook/assets/recipes/mcp-installation.md), [token optimization](plugins/aidd-context/skills/12-cook/assets/recipes/token-optimization.md), [install AIDDy](plugins/aidd-context/skills/12-cook/assets/recipes/install-aiddy-in-codex.md). Project recipes created by cook live in `aidd_docs/recipes/`. |
+| 🍳 **Recipes** | Bundled how-to sheets: [Connect your services](plugins/aidd-context/skills/12-cook/assets/recipes/connect-services.md), [token optimization](plugins/aidd-context/skills/12-cook/assets/recipes/token-optimization.md), [install AIDDy](plugins/aidd-context/skills/12-cook/assets/recipes/install-aiddy-in-codex.md). Project recipes created by cook live in `aidd_docs/recipes/`. |
 | 🏛️ **[Architecture](docs/ARCHITECTURE.md)** | How the framework composes: plugins, skills, hooks, agents. |
 | 🧩 **[Create a plugin](docs/CREATE_PLUGIN.md)** | Build and publish your own. |
 | 🛒 **[Marketplace](docs/MARKETPLACE.md)** | Install scopes and versioning. |

aidd_docs/memory/backlog.md

@@ -5,9 +5,9 @@
 | Support | Authority for | Role |
 | --- | --- | --- |
 | GitHub Issues | the work items, their type and their scope | where a change is agreed before a branch exists |
-| AIDD Roadmap board (org project 8) | order, status, and what ships next | the single source of truth for the roadmap |
+| AIDD Roadmap board (`ai-driven-dev` org project 8) | order, status, and what ships next | the single source of truth for the roadmap |
 | Milestones | themes | one milestone per theme; no milestone means backlog |
-| GitHub Discussions | ideas and their upvotes | where an idea is weighed before it becomes an issue |
+| GitHub Discussions (`ai-driven-dev/framework`) | ideas and their upvotes | where an idea is weighed before it becomes an issue |
 | `aidd_docs/` task documents | the local plans, specs and reviews behind an issue | never a substitute for the issue |
 
 ## Representation

aidd_docs/memory/ecosystem.md

@@ -5,23 +5,19 @@ flowchart LR
   Human([Human])
   Agent([Agent])
   App([App])
-  Tool["AI coding tool · project-brief.md"]
   GitHub["GitHub · vcs.md"]
   Board["Roadmap project board · backlog.md"]
   Discussions["GitHub Discussions · backlog.md"]
   Npm["npm registry · deployment.md"]
   Packages["GitHub Packages · deployment.md"]
-  Discord["Discord"]
+  Discord["Discord · human only"]
   ReleasePlease["release-please · deployment.md"]
   Dependabot["Dependabot · deployment.md"]
 
-  Human -- web --> GitHub
-  Human -- gh --> Board
-  Human -- web --> Discussions
   Human -- web --> Discord
-  Agent -- gh --> GitHub
-  Agent -- gh --> Board
-  Agent -- aidd --> Tool
+  Agent -- cli --> GitHub
+  Agent -- cli --> Board
+  Agent -- cli --> Discussions
   App -- http --> GitHub
   App -- http --> Npm
 

aidd_docs/memory/vcs.md

@@ -8,7 +8,7 @@ The version-control conventions this project follows: branches, commits, and the
 
 - Production branch: `main`. Integration branch: `next`, the target for day-to-day work.
 - GitHub's default branch is `main`, so `gh pr create` without `--base next` targets production.
-- Platform: GitHub, driven through `gh`. Tickets are GitHub Issues.
+- Platform: GitHub, repository `ai-driven-dev/framework`, driven through `gh`. Tickets are GitHub Issues.
 - Pull request template: `.github/PULL_REQUEST_TEMPLATE.md`.
 - The release model — weekly promotion, hotfix path — is in [`RELEASE.md`](../../RELEASE.md); the tooling behind it is in `deployment.md`.
 

aidd_docs/tasks/2026_10/2026_10_14_connect-services/challenge.md

@@ -0,0 +1,28 @@
+**Verdict:** READY
+**Scope:** Service access verification extracted from `02-project-memory` into `aidd-context:13-access-check`, the recipe reworked after the 2026-10-15 review, and onboarding ending its foundations with the check.
+**Evidence limits:** The skill was run by hand against this repository's memory on 2026-10-20, re-run by a fresh session on 2026-10-28; MCP, `http` and `web` probes were not exercised, this bank has `cli` edges only. Client guide URLs checked 2026-10-20.
+
+## Verified
+
+- **Scope:** `02-project-memory/` is identical to `origin/next`; access checking is a separate, read-only skill whose service list comes from `ecosystem.md`, never from the recipe.
+- **Placement:** No file of `13-access-check/` links outside it; the recipe names the skill by slash command only.
+- **Live run:** One row per `Agent` edge of this bank, `Human` and `App` edges skipped:
+
+  | Tool | Mode | Account / target | Status | Evidence / next step |
+  | --- | --- | --- | --- | --- |
+  | GitHub | `cli` | `alexsoyes` / `ai-driven-dev/framework` | ✅ | `gh repo view` read the repo |
+  | Roadmap board | `cli` | `alexsoyes` / `ai-driven-dev` org project 8 | ✅ | `gh project view 8 --owner ai-driven-dev` read "AIDD Roadmap" |
+  | GitHub Discussions | `cli` | `alexsoyes` / `ai-driven-dev/framework` | ✅ | `gh api graphql` read 7 discussions |
+
+  Run 2026-10-28 by a fresh session following the skill, after the bank dropped the edge to the AI host and the `Human` edges its rules drop, added `Agent -- cli --> Discussions`, and `vcs.md` and `backlog.md` named the repository, the board owner and `gh` as the command. No account is recorded: the bank is shared, so any contributor's account passes.
+
+- **Empty bank:** A directory without `aidd_docs/memory/` stops at Require with `/aidd-context:02-project-memory`, no file written.
+- **Recipe:** CLI, MCP, API and SDK each have one example; none targets a service whose CLI the recipe lists; tokens come from the environment; five client MCP guides answer 200.
+
+## Blockers
+
+None.
+
+## Suggestions
+
+None.

aidd_docs/tasks/2026_10/2026_10_14_connect-services/phase-1.md

@@ -0,0 +1,115 @@
+---
+status: done
+---
+
+# Instruction: Create the `13-access-check` skill
+
+## Architecture projection
+
+> Tree of the final files. ✅ create · ✏️ modify · ❌ delete
+
+```txt
+.
+├── docs/
+│   └── CATALOG.md                                   ✏️ row for 13-access-check
+└── plugins/aidd-context/
+    ├── README.md                                    ✏️ row for access-check
+    └── skills/13-access-check/
+        ├── SKILL.md                                 ✅ router, one action
+        ├── actions/
+        │   └── 01-check.md                          ✅ from 02-project-memory/actions/05-verify-connection.md
+        └── assets/
+            └── access-report.md                     ✅ from 02-project-memory/assets/connection-report.md, one row per edge
+```
+
+## User Journey
+
+```mermaid
+flowchart TD
+  A[/aidd-context:13-access-check/] --> B{aidd_docs/memory/ holds a .md?}
+  B -- no --> C[propose 02-project-memory setup, stop]
+  B -- yes --> D{ecosystem.md exists?}
+  D -- no --> E[propose 02-project-memory refresh, stop]
+  D -- yes --> F{any Agent edge?}
+  F -- no --> G[report nothing to check, stop]
+  F -- yes --> H[map each edge to tool, mode, account, target]
+  H --> I[probe one read per edge]
+  I --> J[print the access report]
+```
+
+## Test Scope
+
+```mermaid
+---
+title: Test scope
+---
+journey
+  section Setup
+    Project with memory and ecosystem.md holding Agent edges => bank ready: 5: system
+  section Happy path
+    Run the skill => one report row per Agent edge with tool, mode, account, target, status and evidence: 5: cli
+    Run the skill => no file under the project changed: 5: cli
+  section Edge case - no bank
+    aidd_docs/memory/ empty => run the skill => proposes 02-project-memory setup and stops: 1: cli
+  section Edge case - no ecosystem
+    bank without ecosystem.md => run the skill => proposes 02-project-memory refresh and stops: 1: cli
+  section Edge case - no Agent edge
+    ecosystem.md with App or Human edges only => run the skill => reports nothing to check: 1: cli
+  section Edge case - one access fails
+    one CLI not signed in => run the skill => that row is ❌ with one corrective action, other rows still probed: 1: cli
+  section Edge case - wrong target
+    authenticated on another account => run the skill => that row is ⚠️ naming expected and observed: 1: cli
+```
+
+## Tasks to do
+
+### `1)` Router
+
+> A skill whose name and description say what it checks and what it never does.
+
+1. `SKILL.md` frontmatter: `name: 13-access-check`, description "Check that the assistant can reach the services recorded in project memory, read-only. Use when the user wants to verify or diagnose service access. Not for installing or configuring services.", `argument-hint: tool`.
+2. Flowchart of the journey above, one action `check`.
+3. Transversal rules: read-only (no writes, installs, logins, configuration or memory changes); never print a secret; a configured tool never proves access, only a read matching the intended account and target does; continue past a failure.
+
+### `2)` Check action
+
+> Map the memory to accesses, probe each once, report.
+
+1. `## Input`: `aidd_docs/memory/ecosystem.md`, plus `vcs.md` and `backlog.md` for accounts and targets. No recipe.
+2. `## Process`: **Require** (the three stop cases and their proposal), **Map** (one row per `Agent -- mode --> Tool` edge; an argument narrows to one tool), **Probe** (`cli` => a read-only command, `mcp` => a read tool, `http` => one authenticated GET with the token from the environment), **Report** (fill `access-report.md`).
+3. Status per row: ✅ read matches account and target; ❌ read failed, with one corrective action; 🔌 tool or server absent; ⚠️ authenticated but account or target different (naming expected and observed) or not recorded in memory (next step: record it in memory). `Human` and `App` edges get no row.
+4. `## Test` table covering every case of the test scope.
+
+### `3)` Report asset
+
+> A report whose rows follow the project, not a fixed category list.
+
+1. Columns: Tool, Mode, Account / target, Status, Evidence / next step.
+2. Keep the status legend from `connection-report.md`, drop ⏭️ (an edge in memory is never "not needed").
+3. Inline comment contract, removed from the printed report.
+
+### `4)` Listings
+
+> The skill is discoverable where skills are listed by hand.
+
+1. Add `13-access-check` to the aidd-context table of `docs/CATALOG.md` with action `01-check`.
+2. Add the `access-check` row to `plugins/aidd-context/README.md`.
+
+### `5)` Onboard
+
+> A first setup ends with the assistant's access proven.
+
+1. `00-onboard` `zones.md`: a `service access` foundation after `memory wiring`, command `aidd-context:13-access-check`, met unless memory or wiring was set up this session and the check has not run or been skipped since.
+2. Foundation counts follow: `flow.md`, `screen-map.md`, `03-present.md` test.
+
+## Test acceptance criteria
+
+| Task | Acceptance criteria                                                                                     |
+| ---- | ------------------------------------------------------------------------------------------------------- |
+| 1    | the skill frontmatter passes the pre-commit skill frontmatter and argument-hint checks                   |
+| 1    | no file of the skill links outside `13-access-check/`                                                    |
+| 2    | each stop case names the exact command to run next and changes no file                                   |
+| 2    | a run against this repository lists the `gh` edges to GitHub and the Roadmap board, each with evidence   |
+| 3    | the report has one row per Agent edge and no category row                                                |
+| 4    | `docs/CATALOG.md` and the plugin README both list the skill; regenerated `CATALOG.md` and README counts show 14 aidd-context skills |
+| 5    | an existing repo renders memory as step 1 of 3, and a session that found memory wired lists no access step |

aidd_docs/tasks/2026_10/2026_10_14_connect-services/phase-2.md

@@ -0,0 +1,57 @@
+---
+status: done
+---
+
+# Instruction: Restore `02-project-memory` to memory only
+
+## Architecture projection
+
+> Tree of the final files. ✅ create · ✏️ modify · ❌ delete
+
+```txt
+.
+└── plugins/aidd-context/skills/02-project-memory/
+    ├── SKILL.md                              ✏️ back to origin/next
+    ├── actions/
+    │   └── 05-verify-connection.md           ❌ moved to 13-access-check
+    └── assets/
+        └── connection-report.md              ❌ moved to 13-access-check
+```
+
+## User Journey
+
+```mermaid
+flowchart TD
+  A[/aidd-context:02-project-memory/] --> B[setup, refresh or rewire]
+  B --> C[no verify entry offered]
+```
+
+## Test Scope
+
+```mermaid
+---
+title: Test scope
+---
+journey
+  section Happy path
+    Diff 02-project-memory against origin/next => no difference: 5: cli
+    Grep the repository for 05-verify-connection and connection-report.md => no match: 5: cli
+```
+
+## Tasks to do
+
+### `1)` Revert the skill
+
+> The memory skill does one thing again.
+
+1. Restore `02-project-memory/SKILL.md` from `origin/next`.
+2. Delete `actions/05-verify-connection.md` and `assets/connection-report.md`.
+3. Replace every remaining reference to them (recipe, task docs) with `/aidd-context:13-access-check`; the recipe ones land in phase 3.
+
+## Test acceptance criteria
+
+| Task | Acceptance criteria                                                                    |
+| ---- | -------------------------------------------------------------------------------------- |
+| 1    | `02-project-memory/` is identical to `origin/next`                                     |
+| 1    | the argument hint is `setup \| refresh \| rewire` and the description names no verification |
+| 1    | no file in the repository names `05-verify-connection` or `connection-report.md`       |

aidd_docs/tasks/2026_10/2026_10_14_connect-services/phase-3.md

@@ -0,0 +1,78 @@
+---
+status: done
+---
+
+# Instruction: Apply the review to `connect-services.md`
+
+## Architecture projection
+
+> Tree of the final files. ✅ create · ✏️ modify · ❌ delete
+
+```txt
+.
+└── plugins/aidd-context/skills/12-cook/assets/recipes/
+    └── connect-services.md                   ✏️ intro, step 2 examples, step 3 target
+```
+
+## User Journey
+
+```mermaid
+flowchart TD
+  A[Read the recipe] --> B[record services in project memory]
+  B --> C[choose CLI, MCP, API or SDK per service]
+  C --> D[install and sign in, with the client's own MCP guide]
+  D --> E[/aidd-context:13-access-check/]
+```
+
+## Test Scope
+
+```mermaid
+---
+title: Test scope
+---
+journey
+  section Happy path
+    Validate bundled recipes => connect-services.md has no finding: 5: cli
+    Check markdown links => every link of the recipe resolves: 5: cli
+    Read step 2 => one example each for CLI, MCP, API and SDK: 5: cli
+    Read step 3 => it names /aidd-context:13-access-check and no 02-project-memory action: 5: cli
+```
+
+## Tasks to do
+
+### `1)` Intro
+
+> The reader records services in memory before connecting them (review #7).
+
+1. Add one line: project memory records which services the team uses, in `ecosystem.md`, built by `/aidd-context:02-project-memory`.
+
+### `2)` Step 2 examples
+
+> Four options announced, four shown, none contradicting CLI-first (review #3, #5, #6).
+
+1. Replace the GitHub API example with an internal team API: one `curl --fail` GET, token read from an environment variable, never pasted.
+2. Add an SDK example called from application code (the `App` actor), one official client, a few lines.
+3. One sentence: when a service has neither CLI nor MCP, run `/aidd-context:04-skill-generate` starting from its API.
+
+### `3)` Client coverage
+
+> Each client finds its own MCP setup without five copies of the same config (review #4).
+
+1. Keep the Claude Code `/mcp` example.
+2. Add a table linking the official MCP setup page of Claude Code, Codex, Cursor, GitHub Copilot and OpenCode; fetch each URL before writing it.
+
+### `4)` Step 3
+
+> Verification is one skill call.
+
+1. Replace the two links to `02-project-memory` files and the raw prompt with `/aidd-context:13-access-check`, by slash command, the way `start-a-project.md` names skills.
+
+## Test acceptance criteria
+
+| Task | Acceptance criteria                                                                         |
+| ---- | ------------------------------------------------------------------------------------------- |
+| 1    | the intro names `ecosystem.md` and `/aidd-context:02-project-memory`                        |
+| 2    | no API or SDK example targets a service whose CLI the recipe already lists                  |
+| 2    | no example prints or embeds a token value                                                    |
+| 3    | the client table has five rows, each URL answering 200 on the day it was written            |
+| 4    | step 3 contains `/aidd-context:13-access-check` and no link into `02-project-memory/`       |