rubyn-code 0.3.0 → 0.5.0

data/skills/rubyn_self_test.md

@@ -98,11 +98,98 @@ Score: 18/22 (82%) — 4 failures
 ### 11. Slash Commands (report only — don't execute)
 - Report which slash commands are registered by reading `lib/rubyn_code/cli/commands/registry.rb` or the help output. PASS if at least 15 commands found.
 
-### 12. Architecture Integrity
+### 12. MCP Integration
+- **grep**: Search for `url:.*server_def` in `lib/rubyn_code/mcp/config.rb`. PASS if at least 1 match found (confirms SSE url is extracted — a critical bug was shipped without this).
+- **grep**: Search for `autoload.*Mcp` in `lib/rubyn_code.rb`. PASS if found (confirms `/mcp` command is wired up).
+- **run_specs**: Run `bundle exec rspec spec/rubyn_code/mcp/config_spec.rb --format progress`. PASS if output contains `0 failures`.
+- **bash**: Check if `.rubyn-code/mcp.json` exists in the project root. PASS if exists, SKIP if not (MCP is optional per-project).
+
+### 13. GOLEM Autonomous Mode
+- **grep**: Search for `class Daemon` in `lib/rubyn_code/autonomous/daemon.rb`. PASS if found (confirms daemon framework exists).
+- **grep**: Search for `failed\?` in `lib/rubyn_code/tasks/models.rb`. PASS if found (confirms failed task status is available).
+- **grep**: Search for `total_cost` in `lib/rubyn_code/autonomous/daemon.rb`. PASS if at least 2 matches (confirms cost tracking is implemented).
+- **run_specs**: Run `bundle exec rspec spec/rubyn_code/autonomous/daemon_spec.rb --format progress`. PASS if output contains `0 failures` (verifies lifecycle, retries, cost limits, audit trails, concurrent claiming).
+
+### 14. Architecture Integrity
 - **grep**: Search for `autoload` in `lib/rubyn_code.rb`. PASS if at least 40 autoload entries found.
 - **glob**: Check that all 16 layer directories exist under `lib/rubyn_code/`. PASS if at least 14 found.
 - **read_file**: Read `lib/rubyn_code.rb` and verify it has modules for Agent, Tools, Context, Skills, Memory, Observability, Learning. PASS if all 7 found.
 
+### 15. Skill-Pack Autoload — Live Registry Roundtrip
+
+End-to-end exercise of the autoload pipeline against the real registry at `rubyn.ai`: fetch the catalog, install a pack the user does **not** already have, verify it works on disk + in the catalog + through the matcher, then remove it. The user's pre-existing installed packs are not touched.
+
+- **Live roundtrip**: `bash` with the script below. PASS if the final line is `ROUNDTRIP: PASS`. SKIP if the final line starts with `SKIP:` (means every registry pack is already installed locally — rare).
+
+  ```bash
+  bundle exec ruby -e '
+    require_relative "lib/rubyn_code"
+
+    client = RubynCode::Skills::RegistryClient.new
+    pack_manager = RubynCode::Skills::PackManager.new
+
+    catalog_packs = client.fetch_catalog[:data] || []
+    abort "FAIL: empty registry catalog" if catalog_packs.empty?
+
+    already_installed = catalog_packs.map { |p| p[:name] || p["name"] }
+                                     .select { |n| pack_manager.installed?(n) }
+    target = catalog_packs.find { |p| !already_installed.include?(p[:name] || p["name"]) }
+    abort "SKIP: every registry pack is already installed" if target.nil?
+
+    name = target[:name] || target["name"]
+    puts "TEST PACK: #{name}"
+
+    success = false
+    begin
+      result = client.fetch_pack(name)
+      pack_manager.install(result[:data], etag: result[:etag])
+      raise "install did not create directory" unless pack_manager.installed?(name)
+      puts "STEP install: PASS"
+
+      catalog_obj = RubynCode::Skills::Catalog.new(
+        File.join(Dir.home, ".rubyn-code", "skill-packs")
+      )
+      skills = catalog_obj.available.select { |e| e[:path].include?("/#{name}/") }
+      raise "catalog sees no skills" if skills.empty?
+      raise "no skills have triggers" if skills.none? { |s| !s[:triggers].empty? }
+      puts "STEP catalog: PASS (#{skills.size} skills, #{skills.count { |s| !s[:triggers].empty? }} with triggers)"
+
+      sample = skills.find { |s| !s[:triggers].empty? }
+      matcher = RubynCode::Skills::Matcher.new(catalog: catalog_obj)
+      hits = matcher.match("question about #{sample[:triggers].first}")
+      raise "matcher did not hit on \"#{sample[:triggers].first}\"" if hits.none? { |h| h[:name] == sample[:name] }
+      puts "STEP matcher: PASS (hit \"#{sample[:name]}\" via \"#{sample[:triggers].first}\")"
+      success = true
+    ensure
+      pack_manager.remove(name)
+      if pack_manager.installed?(name)
+        puts "STEP cleanup: FAIL (pack still on disk)"
+      else
+        puts "STEP cleanup: PASS"
+      end
+    end
+
+    puts(success ? "ROUNDTRIP: PASS" : "ROUNDTRIP: FAIL")
+  '
+  ```
+
+  The script:
+  1. Fetches the registry catalog (proves the static API at `rubyn.ai/api/v1/skills/packs.json` is reachable).
+  2. Picks the first pack the user does **not** have installed, so we never touch their existing packs.
+  3. Installs the pack via `RegistryClient#fetch_pack` + `PackManager#install` — exactly the path the autoload pipeline uses.
+  4. Reads the disk back through a fresh `Skills::Catalog` and confirms the new skills are visible with parsed triggers.
+  5. Runs `Skills::Matcher#match` against a real trigger from one of the freshly-installed skills, confirming the matcher would have fired in a real session.
+  6. Removes the pack in an `ensure` block so a partial failure still leaves the user's system clean.
+
+  PASS criteria: `STEP install`, `STEP catalog`, `STEP matcher`, and `STEP cleanup` all `PASS`, with a final `ROUNDTRIP: PASS`.
+
+- **Live autoload notification (manual verification, not scored)**: This isn't a tool call — it's a hint to surface in the scorecard for the user. Tell them: after the self-test completes, send a follow-up prompt that includes a trigger word from any pack in the registry (e.g. `"explain turbo drive"`). The Rubyn renderer should print:
+  ```
+  📥 Fetching skill pack 'hotwire' from registry…
+  📚 Loaded: turbo-drive
+  ```
+  before the next response (the `📥` line appears only if the pack wasn't already installed). Do **not** count this as PASS/FAIL — just mention it in the scorecard so the user can verify the renderer side themselves.
+
 ## Scoring
 
 Count total PASS results out of total tests run. Report the percentage.

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rubyn-code
 version: !ruby/object:Gem::Version
-  version: 0.3.0
+  version: 0.5.0
 platform: ruby
 authors:
 - fadedmaturity
@@ -183,6 +183,7 @@ files:
 - db/migrations/010_create_instincts.sql
 - db/migrations/011_fix_mailbox_messages_columns.rb
 - db/migrations/012_expand_mailbox_message_types.rb
+- db/migrations/013_add_failed_status_to_tasks.rb
 - exe/rubyn-code
 - lib/rubyn_code.rb
 - lib/rubyn_code/agent/RUBYN.md
@@ -200,6 +201,7 @@ files:
 - lib/rubyn_code/agent/tool_processor.rb
 - lib/rubyn_code/agent/usage_tracker.rb
 - lib/rubyn_code/auth/RUBYN.md
+- lib/rubyn_code/auth/key_encryption.rb
 - lib/rubyn_code/auth/oauth.rb
 - lib/rubyn_code/auth/server.rb
 - lib/rubyn_code/auth/token_store.rb
@@ -224,20 +226,27 @@ files:
 - lib/rubyn_code/cli/commands/diff.rb
 - lib/rubyn_code/cli/commands/doctor.rb
 - lib/rubyn_code/cli/commands/help.rb
+- lib/rubyn_code/cli/commands/install_skills.rb
+- lib/rubyn_code/cli/commands/list_skills.rb
+- lib/rubyn_code/cli/commands/mcp.rb
 - lib/rubyn_code/cli/commands/model.rb
 - lib/rubyn_code/cli/commands/new_session.rb
 - lib/rubyn_code/cli/commands/plan.rb
+- lib/rubyn_code/cli/commands/provider.rb
 - lib/rubyn_code/cli/commands/quit.rb
 - lib/rubyn_code/cli/commands/registry.rb
+- lib/rubyn_code/cli/commands/remove_skills.rb
 - lib/rubyn_code/cli/commands/resume.rb
 - lib/rubyn_code/cli/commands/review.rb
 - lib/rubyn_code/cli/commands/skill.rb
+- lib/rubyn_code/cli/commands/skills.rb
 - lib/rubyn_code/cli/commands/spawn.rb
 - lib/rubyn_code/cli/commands/tasks.rb
 - lib/rubyn_code/cli/commands/tokens.rb
 - lib/rubyn_code/cli/commands/undo.rb
 - lib/rubyn_code/cli/commands/version.rb
 - lib/rubyn_code/cli/daemon_runner.rb
+- lib/rubyn_code/cli/first_run.rb
 - lib/rubyn_code/cli/input_handler.rb
 - lib/rubyn_code/cli/renderer.rb
 - lib/rubyn_code/cli/repl.rb
@@ -252,7 +261,9 @@ files:
 - lib/rubyn_code/config/defaults.rb
 - lib/rubyn_code/config/project_config.rb
 - lib/rubyn_code/config/project_profile.rb
+- lib/rubyn_code/config/schema.json
 - lib/rubyn_code/config/settings.rb
+- lib/rubyn_code/config/validator.rb
 - lib/rubyn_code/context/RUBYN.md
 - lib/rubyn_code/context/auto_compact.rb
 - lib/rubyn_code/context/compactor.rb
@@ -273,6 +284,25 @@ files:
 - lib/rubyn_code/hooks/registry.rb
 - lib/rubyn_code/hooks/runner.rb
 - lib/rubyn_code/hooks/user_hooks.rb
+- lib/rubyn_code/ide/adapters/tool_output.rb
+- lib/rubyn_code/ide/client.rb
+- lib/rubyn_code/ide/handlers.rb
+- lib/rubyn_code/ide/handlers/accept_edit_handler.rb
+- lib/rubyn_code/ide/handlers/approve_tool_use_handler.rb
+- lib/rubyn_code/ide/handlers/cancel_handler.rb
+- lib/rubyn_code/ide/handlers/config_get_handler.rb
+- lib/rubyn_code/ide/handlers/config_set_handler.rb
+- lib/rubyn_code/ide/handlers/initialize_handler.rb
+- lib/rubyn_code/ide/handlers/models_list_handler.rb
+- lib/rubyn_code/ide/handlers/prompt_handler.rb
+- lib/rubyn_code/ide/handlers/review_handler.rb
+- lib/rubyn_code/ide/handlers/session_fork_handler.rb
+- lib/rubyn_code/ide/handlers/session_list_handler.rb
+- lib/rubyn_code/ide/handlers/session_reset_handler.rb
+- lib/rubyn_code/ide/handlers/session_resume_handler.rb
+- lib/rubyn_code/ide/handlers/shutdown_handler.rb
+- lib/rubyn_code/ide/protocol.rb
+- lib/rubyn_code/ide/server.rb
 - lib/rubyn_code/index/codebase_index.rb
 - lib/rubyn_code/learning/RUBYN.md
 - lib/rubyn_code/learning/extractor.rb
@@ -281,6 +311,7 @@ files:
 - lib/rubyn_code/learning/shortcut.rb
 - lib/rubyn_code/llm/RUBYN.md
 - lib/rubyn_code/llm/adapters/anthropic.rb
+- lib/rubyn_code/llm/adapters/anthropic_compatible.rb
 - lib/rubyn_code/llm/adapters/anthropic_streaming.rb
 - lib/rubyn_code/llm/adapters/base.rb
 - lib/rubyn_code/llm/adapters/json_parsing.rb
@@ -324,10 +355,19 @@ files:
 - lib/rubyn_code/protocols/interrupt_handler.rb
 - lib/rubyn_code/protocols/plan_approval.rb
 - lib/rubyn_code/protocols/shutdown_handshake.rb
+- lib/rubyn_code/self_test.rb
 - lib/rubyn_code/skills/RUBYN.md
+- lib/rubyn_code/skills/auto_suggest.rb
 - lib/rubyn_code/skills/catalog.rb
 - lib/rubyn_code/skills/document.rb
+- lib/rubyn_code/skills/gemfile_parser.rb
 - lib/rubyn_code/skills/loader.rb
+- lib/rubyn_code/skills/matcher.rb
+- lib/rubyn_code/skills/pack_context.rb
+- lib/rubyn_code/skills/pack_installer.rb
+- lib/rubyn_code/skills/pack_manager.rb
+- lib/rubyn_code/skills/registry_autoload.rb
+- lib/rubyn_code/skills/registry_client.rb
 - lib/rubyn_code/skills/ttl_manager.rb
 - lib/rubyn_code/sub_agents/RUBYN.md
 - lib/rubyn_code/sub_agents/runner.rb
@@ -358,6 +398,8 @@ files:
 - lib/rubyn_code/tools/git_status.rb
 - lib/rubyn_code/tools/glob.rb
 - lib/rubyn_code/tools/grep.rb
+- lib/rubyn_code/tools/ide_diagnostics.rb
+- lib/rubyn_code/tools/ide_symbols.rb
 - lib/rubyn_code/tools/load_skill.rb
 - lib/rubyn_code/tools/memory_search.rb
 - lib/rubyn_code/tools/memory_write.rb