anvil-engineering-workbench 1.0.5 → 1.0.6

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,22 @@
1
1
  # ANVIL Changelog
2
2
 
3
+ ## 1.0.6 — AI Harness MCP Wizard & Dashboard Project Memory Hub
4
+
5
+ ANVIL 1.0.6 introduces automated AI harness integration for modern editors (Cursor, Claude Code, Windsurf, VS Code), a dedicated Project Memory & Architectural Decisions hub, and native dual-path memory retention.
6
+
7
+ ### Features & Capabilities
8
+ - **Automated AI Harness MCP Setup (`anvil mcp setup`):** Single-command configuration for developer environments:
9
+ - `--client cursor`: Generates and non-destructively merges into `.cursor/mcp.json`.
10
+ - `--client claude`: Generates the exact stdio command for Claude Code CLI registration.
11
+ - `--client windsurf` & `--client vscode`: Configures `.windsurf/mcp_config.json` and `.vscode/mcp.json`.
12
+ - `--client all`: Configures all supported IDE harnesses in one execution.
13
+ - **Project Memory & Architectural Decisions Hub (`anvil memory` & Web UI):**
14
+ - Live interactive Project Memory panel directly in the web dashboard under Overview & Health.
15
+ - Dropdown categorization (`decision`, `architecture`, `constraint`, `preference`, `fact`), input tags, and single-click addition/removal.
16
+ - Seamless dual-path storage supporting modern `.anvil/memory.json` with fallback to `.forge/memory.json`.
17
+ - Real-time token-budgeted context ranking so agent instructions receive prioritized architectural constraints without exceeding window limits.
18
+ - **Interactive Landing Page Demos:** Enhanced product site with interactive terminal simulation tabs (`05 mcp-setup`, `06 memory`) and stage explorer showcasing harness connections.
19
+
3
20
  ## 1.0.5 — Seamless Single-Page Dashboard & Showcase Telemetry Hub
4
21
 
5
22
  ANVIL 1.0.5 enhances the local engineering experience with a seamless single-page dashboard architecture, pre-populated interactive showcase telemetry, automatic port fallback (`4173..4183`), and complete release pipeline synchronization.
package/README.md CHANGED
@@ -62,7 +62,7 @@
62
62
  ---
63
63
 
64
64
  > [!NOTE]
65
- > **ANVIL v1.0.5 — AI harness foundation.** The local CLI, local dashboard, agent orchestration, MCP server, provider adapters, GitHub Action, automated self-healing cleaner, byte-accurate rollback (`anvil undo`), pre-commit hooks (`anvil hook install`), and GitHub App installation/authentication foundation are available. See [Current capabilities and boundaries](#current-capabilities-and-boundaries).
65
+ > **ANVIL v1.0.6 — AI harness foundation.** The local CLI, local dashboard, agent orchestration, MCP server, provider adapters, GitHub Action, automated self-healing cleaner, byte-accurate rollback (`anvil undo`), pre-commit hooks (`anvil hook install`), AI harness wizard (`anvil mcp setup`), and GitHub App installation/authentication foundation are available. See [Current capabilities and boundaries](#current-capabilities-and-boundaries).
66
66
 
67
67
  ## What is ANVIL?
68
68
 
@@ -117,7 +117,7 @@ npm install --global anvil-engineering-workbench
117
117
  Or install directly from the GitHub repository:
118
118
 
119
119
  ```bash
120
- npm install --global git+https://github.com/anvil-workbench/ANVIL.git#v1.0.5
120
+ npm install --global git+https://github.com/anvil-workbench/ANVIL.git#v1.0.6
121
121
  ```
122
122
 
123
123
  Verify the installation:
@@ -158,7 +158,7 @@ npm --version
158
158
  git --version
159
159
 
160
160
  # Install ANVIL
161
- npm install --global git+https://github.com/anvil-workbench/ANVIL.git#v1.0.5
161
+ npm install --global git+https://github.com/anvil-workbench/ANVIL.git#v1.0.6
162
162
 
163
163
  # Move into your project
164
164
  cd C:\Users\YourName\Projects\my-project
@@ -538,9 +538,9 @@ jobs:
538
538
  runs-on: ubuntu-latest
539
539
  steps:
540
540
  - uses: actions/checkout@v4
541
- - uses: anvil-workbench/ANVIL@v1.0.5
541
+ - uses: anvil-workbench/ANVIL@v1.0.6
542
542
  with:
543
- version: v1.0.5
543
+ version: v1.0.6
544
544
  project-path: .
545
545
  operation: inspect
546
546
  ```
@@ -744,7 +744,7 @@ Read [CONTRIBUTING.md](CONTRIBUTING.md) before opening a pull request.
744
744
 
745
745
  ## Project history
746
746
 
747
- ANVIL's v1.0.5 release brings seamless single-page dashboard execution, rich pre-populated showcase telemetry, automatic port fallback (`4173..4183`), single-command byte-accurate rollback (`anvil undo`), automated test verification gate (`--verify`), Git pre-commit hook automation (`anvil hook install`), framework-aware entrypoint resolution, full multi-language architecture mapping, automated farzi-file scrubbing, code auto-repair, one-command doctor self-healing, the local CLI, engineering intelligence, agent workflows, MCP integration, GitHub Action and GitHub App installation foundation into one versioned product.
747
+ ANVIL's v1.0.6 release introduces an automated AI harness MCP setup wizard (`anvil mcp setup --client cursor|claude|windsurf|all`), a durable Project Memory & Architectural Decisions hub (`anvil memory` and live dashboard panel), dual-path memory retention (`.anvil/memory.json` / `.forge/memory.json`), seamless single-page dashboard execution, rich pre-populated showcase telemetry, automatic port fallback (`4173..4183`), single-command byte-accurate rollback (`anvil undo`), automated test verification gate (`--verify`), Git pre-commit hook automation (`anvil hook install`), framework-aware entrypoint resolution, full multi-language architecture mapping, automated farzi-file scrubbing, code auto-repair, one-command doctor self-healing, the local CLI, engineering intelligence, agent workflows, MCP integration, GitHub Action and GitHub App installation foundation into one versioned product.
748
748
 
749
749
  The project is maintained as an original ANVIL codebase. ECC is a capability and documentation reference; ANVIL maintains its own identity, implementation, data models and roadmap.
750
750
 
package/dist/cli.js CHANGED
@@ -45,6 +45,7 @@ import { captureIntelligenceSnapshot, compareLatestIntelligence, formatIntellige
45
45
  import { analyzeDependencies, formatDependencyIntelligence } from "./core/dependency-intelligence.js";
46
46
  import { generateRemediationPlan, formatRemediationPlan } from "./core/remediation-planner.js";
47
47
  import { addProjectMemory, buildProjectMemoryContext, listProjectMemory, removeProjectMemory } from "./core/project-memory.js";
48
+ import { setupHarnessMcp, resolveMcpServerScript } from "./core/mcp-setup.js";
48
49
  import { formatSkillContext, loadSkillRegistry, selectSkills } from "./core/skill-registry.js";
49
50
  import { applyBeforeAgentHooks, loadHooks } from "./core/hook-engine.js";
50
51
  import { assertSupportedAgentPolicy, loadPermissionPolicy } from "./core/permission-policy.js";
@@ -125,7 +126,7 @@ function readAgentDurationBudgets() {
125
126
  }
126
127
  return budgets;
127
128
  }
128
- const VERSION = "1.0.5";
129
+ const VERSION = "1.0.6";
129
130
  const HELP = `
130
131
  ANVIL — The AI Engineering Workbench
131
132
 
@@ -184,6 +185,7 @@ Commands:
184
185
  map [path] Map source modules and import relationships
185
186
  agents [path] List built-in and workspace agent roles
186
187
  memory <action> Manage local project memory (add, list, remove, context)
188
+ mcp [setup] Model Context Protocol server and harness setup (cursor, claude, windsurf)
187
189
  skills [path] List built-in and workspace skills
188
190
  hooks [path] List configured declarative lifecycle hooks
189
191
  agent-security [path] Audit agent instructions, skills, hooks, and permissions
@@ -916,9 +918,14 @@ async function main(args) {
916
918
  }
917
919
  case "memory": {
918
920
  const [action = "list", ...memoryArgs] = rest;
919
- const targetIndex = memoryArgs.findIndex(arg => !arg.startsWith("--"));
920
- const target = path.resolve(targetIndex >= 0 ? memoryArgs[targetIndex] ?? "." : ".");
921
921
  const option = (name) => { const index = memoryArgs.indexOf(name); return index >= 0 ? memoryArgs[index + 1] : undefined; };
922
+ const skipValues = new Set();
923
+ for (let i = 0; i < memoryArgs.length; i++) {
924
+ if (memoryArgs[i]?.startsWith("--"))
925
+ skipValues.add(i + 1);
926
+ }
927
+ const positional = memoryArgs.filter((arg, i) => !arg.startsWith("--") && !skipValues.has(i));
928
+ const target = path.resolve(positional[0] ?? ".");
922
929
  try {
923
930
  if (action === "list") {
924
931
  const entries = await listProjectMemory(target);
@@ -932,7 +939,7 @@ async function main(args) {
932
939
  throw new Error("Usage: anvil memory add [path] --kind decision|architecture|constraint|preference|fact --content TEXT [--tags a,b]");
933
940
  const tags = option("--tags")?.split(",") ?? [];
934
941
  const entry = await addProjectMemory(target, { kind, content, tags });
935
- process.stdout.write(`Saved ${entry.id} to .forge/memory.json\n`);
942
+ process.stdout.write(`Saved ${entry.id} to project memory\n`);
936
943
  }
937
944
  else if (action === "remove") {
938
945
  const id = memoryArgs.find(arg => arg.startsWith("mem-"));
@@ -951,7 +958,48 @@ async function main(args) {
951
958
  }
952
959
  }
953
960
  catch (error) {
954
- process.stderr.write("FORGE memory command failed: " + (error instanceof Error ? error.message : String(error)) + "\n");
961
+ process.stderr.write("ANVIL memory command failed: " + (error instanceof Error ? error.message : String(error)) + "\n");
962
+ process.exitCode = 1;
963
+ }
964
+ return;
965
+ }
966
+ case "mcp": {
967
+ const [subcommand = "info", ...mcpArgs] = rest;
968
+ const option = (name) => { const index = mcpArgs.indexOf(name); return index >= 0 ? mcpArgs[index + 1] : undefined; };
969
+ const skipValues = new Set();
970
+ for (let i = 0; i < mcpArgs.length; i++) {
971
+ if (mcpArgs[i]?.startsWith("--"))
972
+ skipValues.add(i + 1);
973
+ }
974
+ const positional = mcpArgs.filter((arg, i) => !arg.startsWith("--") && !skipValues.has(i));
975
+ const target = path.resolve(positional[0] ?? ".");
976
+ try {
977
+ if (subcommand === "setup") {
978
+ const client = (option("--client") ?? "all");
979
+ const report = await setupHarnessMcp(target, client);
980
+ process.stdout.write(`\nANVIL Model Context Protocol Harness Setup\n`);
981
+ process.stdout.write(`Project: ${report.projectPath}\n`);
982
+ process.stdout.write(`MCP Server: ${report.serverScriptPath}\n\n`);
983
+ for (const res of report.results) {
984
+ process.stdout.write(`✓ [${res.client.toUpperCase()}]: ${res.details}\n`);
985
+ }
986
+ process.stdout.write(`\nYour AI harness is ready to use ANVIL tools!\n`);
987
+ }
988
+ else if (subcommand === "server" || subcommand === "start") {
989
+ const script = resolveMcpServerScript();
990
+ await import(script);
991
+ }
992
+ else {
993
+ const script = resolveMcpServerScript();
994
+ process.stdout.write(`ANVIL Model Context Protocol (MCP) Server\n\n`);
995
+ process.stdout.write(`Server entrypoint: ${script}\n\n`);
996
+ process.stdout.write(`Commands:\n`);
997
+ process.stdout.write(` anvil mcp setup [--client cursor|claude|windsurf|vscode|all] Configure your AI harness\n`);
998
+ process.stdout.write(` anvil mcp start Run MCP stdio transport\n\n`);
999
+ }
1000
+ }
1001
+ catch (error) {
1002
+ process.stderr.write("ANVIL MCP command failed: " + (error instanceof Error ? error.message : String(error)) + "\n");
955
1003
  process.exitCode = 1;
956
1004
  }
957
1005
  return;