Feat/dr1001 token whitelist wildcard

.agents/skills/design-system/SKILL.md

@@ -0,0 +1,107 @@
+---
+name: design-system
+description: >-
+  DeepRouter visual design system — MANDATORY for any UI / frontend / visual
+  change in web/default. Use whenever you create or edit React components, pages,
+  CSS, Tailwind classes, colors, typography, spacing, buttons, inputs, badges,
+  cards, modals, layout, hero/marketing sections, or any user-visible surface.
+  Loads the canonical color tokens, type scale, component specs, and the hard
+  "do not" rules from docs/DESIGN.md so the change matches the brand instead of
+  drifting into a generic enterprise look.
+---
+
+# DeepRouter Design System (enforced)
+
+**Canonical source:** [`docs/DESIGN.md`](../../../docs/DESIGN.md). Visual board:
+`docs/brand/index.html`. CSS reference: `docs/brand/deeprouter-brand.css`.
+Brand PRD: `docs/DeepRouter-PRD-brand.md`.
+
+> ⚠️ DESIGN.md is layered. **§0–5 is canonical** (Plus Jakarta Sans, 7px radius,
+> the `:root` token map). **§6–9 is "Historical Inspiration"** — it references an
+> older "Camera Plain" font, 6px radius, and negative letter-spacing that
+> **contradict** the canonical part. On any conflict, **§0–5 wins.** Don't copy
+> the Camera Plain / negative-tracking specifics from §7/§9 into production.
+
+This skill is the gate for **any change a user can see** in `web/default/`. Read
+it before writing the component; match these tokens exactly; prefer existing
+Tailwind/theme tokens and the `.dr-*` classes over one-off hex.
+
+## Non-negotiables (the rules people break)
+
+1. **Cream is the canvas, never pure white.** Page background `#F7F4ED`. Raised
+   surfaces (cards, inputs, popovers, modals) use Soft White `#FCFBF8`. A
+   `#FFFFFF` full-page background is wrong.
+2. **Charcoal, not black.** Text & dark buttons use `#1C1C1C`, secondary text `#5F5F5D`.
+3. **Borders, not shadows, contain cards.** Use a `#ECEAE4` border. Avoid heavy box-shadows.
+4. **AI Blue `#2563FF` is an accent, not décor.** Use it for primary AI action,
+   focus ring, selected state, routing lines, charts — never as a large
+   background, gradient, glassy orb, or blue-purple wash.
+5. **Two weights only: 400 (body/UI) and 600 (headings).** No 700/bold. Hierarchy
+   comes from size + spacing, not weight.
+6. **Normal letter-spacing in product UI.** (The negative-tracking advice is from
+   the historical §7/§9 — ignore it for canonical surfaces.)
+7. **Tabular numbers** for metrics, quotas, prices, latency.
+8. **Use theme tokens / `.dr-*` classes, not raw hex** in components where possible.
+9. **Logo is PNG, never redrawn as SVG.** Don't recolor/gradient/shadow the mark.
+   Avatars: DiceBear `notionists` only (line-art), saved locally — never
+   `avataaars`/`bottts`/monogram circles.
+
+## Color tokens (canonical)
+
+| Token | Hex | Use |
+|---|---|---|
+| Cream | `#F7F4ED` | page background |
+| Soft White | `#FCFBF8` | cards, inputs, modals |
+| Charcoal | `#1C1C1C` | primary text, dark buttons, logo black |
+| Muted | `#5F5F5D` | secondary text, captions |
+| Border | `#ECEAE4` | dividers, card borders |
+| AI Blue | `#2563FF` | primary AI action, focus, selected, routing |
+| Stable Green | `#148F5F` | healthy / success |
+| Warning Orange | `#C76812` | warning / quota pressure |
+| Error Red | `#C9362B` | failure / destructive |
+
+Frontend `:root` mapping (DESIGN.md §5):
+
+```css
+--background:#f7f4ed; --foreground:#1c1c1c; --card:#fcfbf8; --card-foreground:#1c1c1c;
+--primary:#1c1c1c; --primary-foreground:#fcfbf8; --accent:#2563ff;
+--muted-foreground:#5f5f5d; --border:#eceae4; --input:#eceae4; --ring:#2563ff;
+```
+
+## Typography
+
+Font: **Plus Jakarta Sans** (stack falls back to `Public Sans`, which is already
+bundled). Type scale: H1 56/64 · H2 40/48 · H3 28/36 · H4 20/24 · Body 16/24 ·
+Small 14/20 · Caption 12/16. Headings weight 600; everything else 400. Dashboard
+labels 12–13px. No oversized hero type inside cards/modals/tables.
+
+## Component specs
+
+- **Buttons** — Primary `#1C1C1C` bg / `#FCFBF8` text · AI Action `#2563FF` bg /
+  white · Secondary `#FCFBF8` / `#1C1C1C` with `rgba(28,28,28,.18)` border · Ghost
+  transparent. Height 40–42px, **radius 7px**, padding 14–18px, font 14/20 weight
+  500–600, focus ring `0 0 0 3px rgba(37,99,255,.14)`.
+- **Inputs** — height 42px, radius 7px, bg `#FCFBF8`, border `rgba(28,28,28,.14)`,
+  focus border `#2563FF` + ring `rgba(37,99,255,.14)`.
+- **Badges** — radius 999px (pill), height 28–30px, font 14 weight 600; tinted
+  bg+text per state (active=blue, beta=charcoal, stable=green, warning=orange, error=red).
+- Pills (999px radius) are for badges / icon-action toggles only — **not** for
+  rectangular buttons (those are 7px).
+
+## `.dr-*` CSS reference classes (in `docs/brand/deeprouter-brand.css`)
+
+`.dr-surface .dr-card .dr-panel` · `.dr-heading-{1,2,3} .dr-body .dr-small
+.dr-caption` · `.dr-button{,-primary,-ai,-secondary,-ghost}` · `.dr-input
+.dr-select` · `.dr-badge{,-active,-beta,-stable,-warning,-error}` · `.dr-metric-card
+.dr-table .dr-sidebar .dr-nav-item .dr-modal`. Use these for prototypes/docs and as
+the spec when building the React equivalents.
+
+## Before you finish a UI change — checklist
+
+- [ ] Background cream `#F7F4ed`, not white; raised surfaces soft-white.
+- [ ] Cards use `#ECEAE4` borders, not box-shadows.
+- [ ] AI Blue only as accent (action/focus/selected/routing), no large gradients/orbs.
+- [ ] Only weights 400 / 600; no bold-700.
+- [ ] Rectangular buttons/inputs at 7px radius; pills only for badges/icon toggles.
+- [ ] Colors come from theme tokens / `.dr-*`, not stray hex.
+- [ ] If it's a customer-facing surface, you ALSO followed CLAUDE.md §0 + the business PRDs (design ≠ exemption from the casual-user rules).

.claude/commands/dr-pr.md

@@ -0,0 +1,32 @@
+Help create a clean PR for the current branch. Follow this checklist before opening the PR.
+
+## Pre-PR checklist
+
+1. **Run tests** — confirm all Go tests pass:
+   ```
+   docker run --rm -v "$(pwd):/app" -w /app golang:1.25-alpine sh -c 'go test ./relay/ ./internal/... 2>&1 | grep -E "^(ok|FAIL)"'
+   ```
+   All lines must say `ok`. If any say `FAIL`, stop and fix before proceeding.
+
+2. **Check diff** — run `git diff main...HEAD` and summarise what changed. Flag any:
+   - Accidental debug prints or TODOs left in
+   - Files that shouldn't be in this PR (seed scripts, .env, temp files)
+   - Missing test for the change
+
+3. **Check ordering bug** — if the PR touches `relay/*_handler.go`, verify that for each handler that calls `applyAirbotixPolicy*`, the call comes **BEFORE** `helper.ModelMappedHelper`. This ordering is critical for kids_mode whitelist correctness.
+
+4. **Push branch** if not already pushed:
+   ```
+   git push -u origin <branch-name>
+   ```
+
+5. **Create the PR via GitHub MCP** — use the create_pull_request tool with:
+   - title: `type(scope): short description` (e.g. `fix(relay): apply policy before model mapping`)
+   - base: `main`
+   - body sections: Problem, Fix, Verification table (test cases + results)
+   - draft: false
+
+## Reminders
+- One concern per PR. Don't bundle unrelated fixes.
+- Co-author line in description: `Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>`
+- After PR is created, share the URL with the user.

.claude/commands/dr-status.md

@@ -0,0 +1,39 @@
+Report the current DeepRouter project status. Do the following steps in order:
+
+1. Run `git log --oneline -8` to show recent commits.
+2. Run `git branch` to list local branches and note any active feature/fix branches.
+3. Run `git status --short` to show any unstaged or uncommitted changes.
+4. Read `AIRBOTIX.md` (the "What we customise" table) to get the Airbotix-specific package status.
+5. Read `PLAN.md` if it exists, and note the current phase.
+
+Then produce a concise report in this format:
+
+---
+## DeepRouter Status — [today's date]
+
+### Recent commits (last 8)
+[list]
+
+### Active branches
+[list any non-main branches]
+
+### Uncommitted changes
+[list or "none"]
+
+### Sprint 1 ticket status (from memory + code)
+| Ticket | Title | Status |
+|--------|-------|--------|
+| DR-6   | internal/billing webhook dispatcher | ✅ Done |
+| DR-7   | internal/kids hard constraints | ✅ Done |
+| DR-8   | internal/policy decision engine | ✅ Done |
+| DR-9   | e2e: same endpoint, different key → different policy | 🟡 PR open, fix incomplete (claude/gemini/responses handlers still have ordering bug) |
+| DR-13  | Quota check RPM/TPM + staging deploy | ⏳ Not started |
+
+### Open PRs / branches
+[describe any open branches/PRs]
+
+### What needs doing next
+[top 1-2 items]
+---
+
+Be specific and honest. Do not mark anything Done if it has known gaps.

.claude/commands/dr-test.md

@@ -0,0 +1,54 @@
+Run the standard DeepRouter policy e2e verification (DR-9 test suite).
+
+The local dev stack runs at http://localhost:3000.
+
+## What you need first
+
+Ask the user for two API tokens if not already provided:
+- ROOT_KEY: a token belonging to a user with `kids_mode=false`, `policy_profile=passthrough`
+- KIDS_KEY: a token belonging to a user with `kids_mode=true`, `policy_profile=kid-safe`
+
+The Groq channel must have `model_mapping`: `gpt-4o-mini` → `llama-3.1-8b-instant`.
+
+## Run 3 test cases
+
+For each test, run the curl command and record the HTTP status + first few words of the response content.
+
+**TEST 1 — root key, non-whitelisted model (should PASS)**
+```
+curl -s -w "\nHTTP %{http_code}" http://localhost:3000/v1/chat/completions \
+  -H "Authorization: Bearer $ROOT_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"model":"llama-3.1-8b-instant","messages":[{"role":"user","content":"Say hello in 5 words"}],"max_tokens":20}'
+```
+Expected: HTTP 200, content with words.
+
+**TEST 2 — kids key, non-whitelisted model (should BLOCK)**
+```
+curl -s -w "\nHTTP %{http_code}" http://localhost:3000/v1/chat/completions \
+  -H "Authorization: Bearer $KIDS_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"model":"llama-3.1-8b-instant","messages":[{"role":"user","content":"Say hello in 5 words"}],"max_tokens":20}'
+```
+Expected: HTTP 400, error mentioning `model_not_eligible_for_kids_mode`.
+
+**TEST 3 — kids key, whitelisted model that maps to non-whitelisted upstream (should PASS)**
+```
+curl -s -w "\nHTTP %{http_code}" http://localhost:3000/v1/chat/completions \
+  -H "Authorization: Bearer $KIDS_KEY" \
+  -H "Content-Type: application/json" \
+  -d '{"model":"gpt-4o-mini","messages":[{"role":"user","content":"Say hello in 5 words"}],"max_tokens":20}'
+```
+Expected: HTTP 200, content with words (channel remaps gpt-4o-mini → llama-3.1-8b-instant internally, but whitelist check sees the original name).
+
+## Report
+
+After running all 3 tests, report a table:
+
+| Test | Key | Model sent | Expected | Result | Pass? |
+|------|-----|------------|----------|--------|-------|
+| 1 | root | llama-3.1-8b-instant | 200 | ... | ✅/❌ |
+| 2 | kids | llama-3.1-8b-instant | 400 | ... | ✅/❌ |
+| 3 | kids | gpt-4o-mini | 200 | ... | ✅/❌ |
+
+If any test fails, diagnose why (check container logs: `docker logs new-api-dev --tail 30`).

.dockerignore

@@ -8,4 +8,6 @@ docs
 .eslintcache
 .gocache
 /web/node_modules
+web/default/node_modules
+web/classic/node_modules
 !THIRD-PARTY-LICENSES.md

.gitattributes

@@ -10,6 +10,7 @@
 *.yml text eol=lf
 *.toml text eol=lf
 *.md text eol=lf
+*.py text eol=lf
 
 # JavaScript/TypeScript files
 *.js text eol=lf
@@ -39,4 +40,4 @@ web/** linguist-vendored
 
 # Un-vendor core frontend source to keep JavaScript visible in language stats
 web/src/components/** linguist-vendored=false
-web/src/pages/** linguist-vendored=false
+web/src/pages/** linguist-vendored=false

.githooks/pre-commit

@@ -0,0 +1,53 @@
+#!/usr/bin/env bash
+# Pre-commit secret scanner — blocks commits containing credential patterns.
+#
+# Patterns caught:
+#   sk-[A-Za-z0-9]{30,}   DeepRouter / OpenAI-style bearer tokens
+#   sk-ant-...             Anthropic API keys
+#   AIzaSy...              Google / Gemini API keys
+#   AKIA...                AWS access key IDs
+#
+# To bypass in a genuine emergency (e.g. committing a test fixture with a
+# clearly-fake token):
+#   git commit --no-verify   ← requires explicit flag, not a default
+
+set -euo pipefail
+
+RED='\033[0;31m'
+YELLOW='\033[1;33m'
+RESET='\033[0m'
+
+# Patterns: each entry is "LABEL|REGEX"
+PATTERNS=(
+  "DeepRouter/OpenAI bearer token|sk-[A-Za-z0-9]{30,}"
+  "Anthropic API key|sk-ant-[A-Za-z0-9\-]{20,}"
+  "Google/Gemini API key|AIzaSy[A-Za-z0-9\-_]{30,}"
+  "AWS access key ID|AKIA[A-Z0-9]{16}"
+)
+
+STAGED=$(git diff --cached --name-only --diff-filter=ACM 2>/dev/null)
+[[ -z "$STAGED" ]] && exit 0
+
+FOUND=0
+for entry in "${PATTERNS[@]}"; do
+  label="${entry%%|*}"
+  regex="${entry##*|}"
+  matches=$(git diff --cached -U0 -- $STAGED 2>/dev/null \
+    | grep '^+' \
+    | grep -v '^+++' \
+    | grep -E "$regex" || true)
+  if [[ -n "$matches" ]]; then
+    echo -e "${RED}[SECRET SCAN] Potential credential blocked: ${label}${RESET}"
+    echo -e "${YELLOW}${matches}${RESET}"
+    FOUND=1
+  fi
+done
+
+if [[ $FOUND -ne 0 ]]; then
+  echo ""
+  echo -e "${RED}Commit blocked. Remove the credential and use an environment variable instead.${RESET}"
+  echo "To bypass (only for clearly fake test fixtures): git commit --no-verify"
+  exit 1
+fi
+
+exit 0