Fix/ruff lint autofixes

.claude/commands/create-spec.md

@@ -457,7 +457,7 @@ Create a new file using this XML structure:
 **Output path:** `$ARGUMENTS/prompts/initializer_prompt.md`
 
 If the output directory has an existing `initializer_prompt.md`, read it and update the feature count.
-If not, copy from `.claude/templates/initializer_prompt.template.md` first, then update.
+If not, copy from `.opencode/templates/initializer_prompt.template.md` first, then update.
 
 **CRITICAL: You MUST update the feature count placeholder:**
 

.env.example

@@ -3,10 +3,14 @@
 
 # CLI Command Selection
 # Choose which CLI command to use for the agent.
-# - claude: Uses Anthropic's official Claude Code CLI (default)
+# - opencode: Uses Opencode's CLI or SDK-backed workflow (default)
 # - glm: Uses GLM CLI (or any other compatible CLI)
-# Defaults to 'claude' if not specified
-# CLI_COMMAND=claude
+# Defaults to 'opencode' if not specified
+# CLI_COMMAND=opencode
+
+# Opencode API Key (preferred):
+# Set OPENCODE_API_KEY to use the Opencode SDK directly instead of a CLI.
+# OPENCODE_API_KEY=your_key_here
 
 # Playwright Browser Mode
 # Controls whether Playwright runs Chrome in headless mode (no visible browser window).

.github/copilot-instructions.md

@@ -0,0 +1,53 @@
+# Copilot Instructions for Contributors (Autocoder)
+
+Purpose: give AI coding agents (and humans) the immediate context and concrete commands needed to be productive in this repo.
+
+Quick-start workflow
+- Python setup: `python -m venv venv && source venv/bin/activate && pip install -r requirements.txt`.
+- Run the CLI agent: `python autonomous_agent_demo.py --project-dir <path|registered_name>`.
+- YOLO (fast prototyping): add `--yolo` (skips Playwright-based browser tests and marks features as passing after lint/type-check).
+- Web UI: in `ui/` run `npm ci` then `npm run dev` (dev) or `npm run build` (production). `start_ui.sh`/`start_ui.bat` serve `ui/dist/`.
+
+Key verification & CI commands
+- Lint: `ruff check .`
+- Security checks: `python test_security.py` (covers bash command allowlist and scripts like `init.sh`).
+- Full local check (as used in project templates): `ruff check . && python test_security.py && cd ui && npm run lint && npm run build`.
+- GitHub CI uses Python 3.11 and Node 20 (.github/workflows/ci.yml).
+
+Big-picture architecture & data flow (short)
+- Agent runtime: `autonomous_agent_demo.py` (entry) -> `agent.py` (session loop) -> `client.py` (Opencode client adapter config, MCP servers, allowed tool lists).
+- MCP servers: `mcp_server/feature_mcp.py` exposes feature management tools; Playwright MCP available for browser automation in non-YOLO mode.
+- Features & tests: stored in project `features.db` (SQLAlchemy models in `api/database.py`). Agents operate via MCP tools like `mcp__features__feature_get_next` and `mcp__features__feature_mark_passing`.
+- UI: FastAPI backend `start_ui.py` + `server/routers/*`; `ui/src/*` consumes REST + WebSocket (`/ws/projects/{project_name}`) for logs and progress.
+
+Project-specific conventions & important patterns
+- Prompt fallback chain: project-specific `prompts/{name}.md` -> `.opencode/templates/{name}.template.md` (See `prompts.py`).
+- YOLO mode: `--yolo` flag in `autonomous_agent_demo.py` and toggle in UI. In YOLO:
+  - Playwright MCP server is skipped
+  - Regression testing is skipped
+  - Features are marked passing on lint/type-check success
+- Security: all bash commands are validated in `security.py` and tested in `test_security.py`. Example rules:
+  - `chmod +x init.sh` is allowed; numeric modes (`chmod 777`) and recursive `-R` are blocked.
+  - `./init.sh` is allowed; `bash init.sh` or other script names may be blocked.
+  - When changing the allowlist, update `test_security.py` accordingly.
+- Tool allowlists and Playwright tooling are configured in `client.py`. Note: `mcp__playwright__browser_run_code` was removed due to crashes—avoid reintroducing without tests.
+
+Where to look first (most impact)
+- `OPENCODE.md` — project overview & workflows (already maintained and authoritative).
+- `client.py` — how agents are configured, allowed tools, and Playwright headless flag (`PLAYWRIGHT_HEADLESS`).
+- `security.py` & `test_security.py` — bash security rules and tests.
+- `mcp_server/feature_mcp.py` — available MCP tools and their behavior.
+- `autonomous_agent_demo.py` / `agent.py` — agent lifecycle and prompts used.
+- `ui/` — run and build steps; `ui/src/hooks/useWebSocket.ts` and `ui/src/hooks/useProjects.ts` for real-time behavior.
+
+Examples agents should follow
+- To run the same checks CI uses: `ruff check . && python test_security.py && cd ui && npm run lint && npm run build`.
+- To run agent locally with Playwright (standard mode): `python autonomous_agent_demo.py --project-dir <project>` (ensure Playwright is available or use default environment from `start.sh`).
+- When adding an MCP tool, add a corresponding test and document it in `mcp_server/feature_mcp.py` and `client.py`.
+
+Editing guidelines
+- Keep prompt changes additive and preserve existing prompt templates in `.opencode/templates`. (rename `.claude/templates` → `.opencode/templates`)
+- If you change security rules or allowed commands, add/adjust tests in `test_security.py`.
+- When updating the UI, remember the `start_ui.*` scripts serve the pre-built `ui/dist/`; a `npm run build` is required before using those scripts.
+
+If anything here is unclear or you'd like a different level of detail (examples or links to specific functions), tell me which sections to expand and I'll iterate. ✅

.claude/agents/code-review.md → .opencode/agents/code-review.md

@@ -123,14 +123,14 @@ Structure your review as follows:
 - Be thorough but constructive - explain why something is an issue
 - Provide specific, actionable feedback with examples
 - Acknowledge good code when you see it
-- Consider the project's existing patterns and conventions (from CLAUDE.md)
+- Consider the project's existing patterns and conventions (from OPENCODE.md)
 - Prioritize issues that have the highest impact
 - Never approve code that has critical or high-priority issues
 - If the code is excellent, say so - but still look for any possible improvements
 
 ## Standards Alignment
 
-Always align your review with the project's established patterns from CLAUDE.md, including:
+Always align your review with the project's established patterns from OPENCODE.md, including:
 - The project's architecture and design patterns
 - Existing coding conventions
 - Technology-specific best practices

.claude/agents/coder.md → .opencode/agents/coder.md

@@ -22,7 +22,7 @@ Before writing ANY code, you MUST:
    - Existing similar implementations to use as reference
    - Configuration files (package.json, pyproject.toml, tsconfig.json, etc.)
    - README files and documentation
-   - CLAUDE.md or similar project instruction files
+   - OPENCODE.md or similar project instruction files
 
 2. **Identify Patterns and Standards**: Search for and document:
    - Naming conventions (files, functions, classes, variables)

.claude/agents/deep-dive.md → .opencode/agents/deep-dive.md

@@ -83,7 +83,7 @@ You have access to powerful tools - USE THEM EXTENSIVELY:
 
 4. **Risk Awareness**: Always consider what could go wrong. Security, performance, maintainability, edge cases.
 
-5. **Context Sensitivity**: Align recommendations with the project's existing patterns, constraints, and standards (including any CLAUDE.md guidance).
+5. **Context Sensitivity**: Align recommendations with the project's existing patterns, constraints, and standards (including any OPENCODE.md guidance).
 
 ## Output Structure
 

.claude/commands/checkpoint.md → .opencode/commands/checkpoint.md

@@ -37,4 +37,4 @@ IMPORTANT:
 - Do NOT skip any files - include everything
 - Make the commit message descriptive enough that someone reviewing the git log can understand what was accomplished
 - Follow the project's existing commit message conventions (check git log first)
-- Include the Claude Code co-author attribution in the commit message
+- Include the Opencode co-author attribution in the commit message