@ryuenn3123/agentic-senior-core 4.3.8 → 4.3.11

package/.agent-context/prompts/bootstrap-design.md

@@ -7,8 +7,24 @@ Use this prompt for UI, UX, frontend layout, screen, or redesign work. Create or
 - Use current repo evidence, project docs, and `.agent-context/` as style context.
 - Do not copy layout rhythm, palette, component skin, or brand posture from external references without explicit user approval.
 - WCAG 2.2 AA is the hard compliance floor.
+- When generating any UI component that renders or accepts user data, explicitly state in the prompt: "ensure all user-supplied input is sanitized, all status indicators have non-color alternatives, and all interactive elements have visible focus states."
 - Before choosing a new UI library, research current official docs.
 
+## Design Reference Retrieval
+
+Before generating tokens from scratch, check if a proven brand reference matches
+the product aesthetic. Run `npx -y getdesign list` to see all available references.
+Pick by spatial mechanics and typographic character, not by color.
+
+Fetch: `npx -y getdesign add <slug> --out docs/DESIGN.md --force`
+
+Customize: replace brand colors with product OKLCH equivalents, replace brand name
+with product name, remove brand iconography, keep spatial rhythm and density logic.
+
+If no brand matches or terminal is unavailable, skip and generate from scratch below.
+
+Source: https://github.com/VoltAgent/awesome-design-md
+
 ## Design Direction Process
 
 1. **Name Your Defaults**: Name three temptations (e.g. SaaS admin default, AI-startup landing) and why they flatten this product. Derive your search direction by rejecting them.
@@ -22,7 +38,7 @@ When the user says "redesign from zero": treat existing UI as behavioral evidenc
 
 ## Output Format
 
-`docs/DESIGN.md` must be a compact token file under 400 tokens. Rationale stays in working memory.
+`docs/DESIGN.md` must be a compact token file under 400 tokens (exception: if using a fetched brand reference from `getdesign`, you may preserve its rich details up to 2500 tokens). Rationale stays in working memory.
 
 Required sections (in this order, no additions):
 
@@ -50,3 +66,13 @@ Only three gates. Do not add more:
 1. Anchor exists and is specific.
 2. Token values are derived from the anchor.
 3. WCAG 2.2 AA compliance floor is met.
+
+## Pre-Commit Critique
+
+Before committing any UI output, answer:
+
+1. Rename test: if the product name changed to an unrelated category, would this UI still look identical? Yes = too generic, reset anchor.
+2. Token drift test: does any color, font, or spacing in the output contradict the DESIGN.md token definitions? Yes = align before commit.
+3. Security test: is user-controlled data rendered unsanitized, or does any status depend on color alone? Yes = fix before merge.
+
+Max 3 revision cycles. If unresolved, reset the anchor entirely.

package/AGENTS.md

@@ -19,7 +19,7 @@ Write instructions as imperative gates:
 
 ## Bootstrap Receipt
 For non-trivial work, run `agentic-senior-core context "<task_description>"` (or `npx @ryuenn3123/agentic-senior-core context "<task_description>"`) and emit a concise Bootstrap Receipt before implementation output or file edits:
-- `loaded_files`: files actually read
+- `loaded_files`: files actually read (exact paths from selectedRules/selectedPrompts only; zero hallucination)
 - `selected_rules`: files selected for this scope and why
 - `skipped_rules`: out-of-scope categories left unloaded
 - `unreachable_files`: required files that could not be read

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ryuenn3123/agentic-senior-core",
-  "version": "4.3.8",
+  "version": "4.3.11",
   "type": "module",
   "description": "Change your AI Agent to code like a Staff Engineer, not a Junior.",
   "bin": {