anvil-engineering-workbench 1.0.0 → 1.0.2
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 +19 -0
- package/README.md +3 -3
- package/SECURITY.md +1 -1
- package/dist/core/dashboard-server.js +6 -6
- package/package.json +2 -2
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,24 @@
|
|
|
1
1
|
# ANVIL Changelog
|
|
2
2
|
|
|
3
|
+
## 1.0.2 — Maintainer & Contact Verification
|
|
4
|
+
|
|
5
|
+
ANVIL 1.0.2 updates official maintainer contact information, aligning all security, support, and package metadata with Aditya Tiwari (`m.g.atiwari28@gmail.com`).
|
|
6
|
+
|
|
7
|
+
### Changes
|
|
8
|
+
- Updated maintainer and support contacts across `README.md`, `SECURITY.md`, and `package.json` to Aditya Tiwari (`m.g.atiwari28@gmail.com`).
|
|
9
|
+
- Refreshed package metadata for npm registry synchronization.
|
|
10
|
+
|
|
11
|
+
## 1.0.1 — Rebranding, Modern Website & Documentation Polish
|
|
12
|
+
|
|
13
|
+
ANVIL 1.0.1 introduces a completely redesigned product website, high-definition SVG iconography, comprehensive documentation unification under the ANVIL brand, and cross-platform stability improvements.
|
|
14
|
+
|
|
15
|
+
### Enhancements & Fixes
|
|
16
|
+
- **Modern Landing Website (`public/index.html`):** Redesigned with an ultra-clean Obsidian and Anvil Copper dark design system, interactive package manager installer tabs (`npm`, `npx`, `pnpm`, `yarn`), interactive live terminal simulation with real-time output switching, and interactive 4-stage workbench explorer.
|
|
17
|
+
- **Self-Contained Vector Iconography:** Replaced external image references with 100% self-contained inline SVG brand marks to ensure zero broken images when viewing locally or offline.
|
|
18
|
+
- **Brand Unification:** Purged all remaining legacy branding across `docs/`, `action.yml`, workflow files, and dashboard server templates.
|
|
19
|
+
- **High-Resolution Banner:** Generated a brand new, crisp screenshot of the ANVIL engineering dashboard (`docs/anvil-dashboard.png`) for repository and npm README presentations.
|
|
20
|
+
- **Cross-Platform Test Stabilization:** Resolved Windows symlink traversal and 8.3 path canonicalization issues in CI workflows.
|
|
21
|
+
|
|
3
22
|
## 1.0.0 — First major release
|
|
4
23
|
|
|
5
24
|
ANVIL 1.0.0 brings the independent ANVIL engineering workbench together as a versioned, installable product. It is designed to provide a local-first alternative for teams and developers who want codebase intelligence, agent orchestration, reviewable workflows, and measurable verification.
|
package/README.md
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
<!-- ANVIL — AI Harness for software engineering -->
|
|
2
2
|
|
|
3
3
|
<p align="center">
|
|
4
|
-
<img src="https://raw.githubusercontent.com/anvil-workbench/ANVIL/main/docs/anvil-dashboard.png" alt="ANVIL AI Harness dashboard" width="100%" />
|
|
4
|
+
<img src="https://raw.githubusercontent.com/anvil-workbench/ANVIL/main/docs/anvil-dashboard.png?v=2" alt="ANVIL AI Harness dashboard" width="100%" />
|
|
5
5
|
</p>
|
|
6
6
|
|
|
7
7
|
<h1 align="center">ANVIL — The AI Engineering Harness</h1>
|
|
@@ -755,11 +755,11 @@ For release-by-release changes, see [CHANGELOG.md](CHANGELOG.md).
|
|
|
755
755
|
|
|
756
756
|
## Support
|
|
757
757
|
|
|
758
|
-
For questions about ANVIL or help with the GitHub App, contact [
|
|
758
|
+
For questions about ANVIL or help with the GitHub App, contact [m.g.atiwari28@gmail.com](mailto:m.g.atiwari28@gmail.com).
|
|
759
759
|
|
|
760
760
|
## Maintainer
|
|
761
761
|
|
|
762
|
-
**
|
|
762
|
+
**Aditya Tiwari** · [@AdityaaaTiwari](https://github.com/AdityaaaTiwari)
|
|
763
763
|
|
|
764
764
|
## License
|
|
765
765
|
|
package/SECURITY.md
CHANGED
|
@@ -4,7 +4,7 @@ ANVIL is an early-stage project. Do not use it to run untrusted automation again
|
|
|
4
4
|
|
|
5
5
|
If you believe you have found a security vulnerability in ANVIL, report it privately before public disclosure.
|
|
6
6
|
|
|
7
|
-
For general ANVIL support, contact [
|
|
7
|
+
For general ANVIL support, contact [m.g.atiwari28@gmail.com](mailto:m.g.atiwari28@gmail.com).
|
|
8
8
|
|
|
9
9
|
## Scope
|
|
10
10
|
|
|
@@ -57,7 +57,7 @@ const PAGE = String.raw `<!doctype html>
|
|
|
57
57
|
<html lang="en">
|
|
58
58
|
<head>
|
|
59
59
|
<meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
|
|
60
|
-
<meta name="theme-color" content="#090b10"><title>ANVIL —
|
|
60
|
+
<meta name="theme-color" content="#090b10"><title>ANVIL — Engineering Workbench</title>
|
|
61
61
|
<style>
|
|
62
62
|
:root{color-scheme:dark;--bg:#090b10;--panel:#11151e;--panel2:#151b27;--line:#252d3b;--muted:#8e9aab;--text:#edf2f8;--accent:#a3e635;--cyan:#67e8f9}
|
|
63
63
|
*{box-sizing:border-box}body{margin:0;background:radial-gradient(ellipse at 75% -15%,#1c2935 0,transparent 38%),var(--bg);color:var(--text);font:14px/1.5 Inter,ui-sans-serif,system-ui,-apple-system,Segoe UI,sans-serif}
|
|
@@ -72,8 +72,8 @@ const PAGE = String.raw `<!doctype html>
|
|
|
72
72
|
<section class="section columns"><div class="card"><div class="section-head"><h2>Specialist agents</h2><span class="pill">AGENT REGISTRY</span></div><div id="agents"><div class="loading">Loading agent roles…</div></div></div><div class="card"><div class="section-head"><h2>Recent runs</h2><span class="pill">LOCAL HISTORY</span></div><div id="runs"><div class="loading">Loading saved runs…</div></div></div><div class="card"><div class="section-head"><h2>Project signals</h2><span class="pill">INSIGHTS</span></div><div id="signals" class="list"><div class="loading">Collecting project signals…</div></div></div></section>
|
|
73
73
|
<section class="section card"><div class="section-head"><h2>Project Health Center</h2><div class="actions"><span class="pill">SNAPSHOT + TRENDS</span><button class="btn" id="refreshHealth">Refresh snapshot</button></div></div><div id="healthPanel" class="list"><div class="loading">Generating project health snapshot…</div></div></section>\n<section class="section columns"><div class="card"><div class="section-head"><h2>Regression intelligence</h2><div class="actions"><span class="pill">VERIFICATION TRENDS</span><button class="btn" id="refreshRegression">Refresh</button></div></div><div id="regressionPanel" class="list"><div class="loading">Comparing saved verification runs…</div></div></div><div class="card"><div class="section-head"><h2>Code review history</h2><div class="actions"><span class="pill">SAVED REVIEWS</span><button class="btn" id="refreshReviews">Refresh</button></div></div><div id="reviewPanel" class="list"><div class="loading">Loading saved review snapshots…</div></div></div></section>
|
|
74
74
|
<section class="section columns"><div class="card"><div class="section-head"><h2>Engineering quality</h2><span class="pill">STATIC AUDIT</span></div><div id="qualityPanel" class="list"><div class="loading">Loading quality findings…</div></div></div><div class="card"><div class="section-head"><h2>Verification history</h2><span class="pill">REAL COMMAND RESULTS</span></div><div id="verificationPanel" class="list"><div class="loading">Loading verification history…</div></div></div></section>\n<section class="section card"><div class="section-head"><h2>Code search</h2><span class="pill">SYMBOL-AWARE SOURCE SEARCH</span></div><div class="actions"><input id="codeSearchQuery" type="search" placeholder="Search symbols or source text…" style="flex:1;min-width:0;background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><button class="btn" id="codeSearchButton">Search</button></div><div id="codeSearchMeta" class="run-meta">Searches source files only; results open in the workspace preview.</div><div id="codeSearchResults" class="list"><div class="empty">Enter at least two characters to search.</div></div></section>\n<section class="section card"><div class="section-head"><h2>Workspace explorer</h2><span class="pill">SAFE FILE PREVIEW</span></div><div class="columns"><div><div class="actions"><button class="btn" id="workspaceUp">↑ Up</button><span id="workspacePath" class="pill">/</span></div><div id="workspaceFiles" class="list" style="margin-top:10px"><div class="loading">Loading workspace files…</div></div></div><div><div id="workspacePreview" class="inspector"><div class="empty">Select a source file to preview it here.</div></div></div></div></section>
|
|
75
|
-
<section class="section card" id="agentConfigStudio"><div class="section-head"><h2>Agent Configuration Studio</h2><span class="pill">WORKSPACE-OWNED MARKDOWN</span></div><p class="run-meta">Edit custom agent and skill instructions in .
|
|
76
|
-
<section class="section card" id="workflowEditor"><div class="section-head"><h2>Visual workflow editor</h2><span class="pill">REUSABLE AGENT PIPELINES</span></div><p class="run-meta">Compose specialist tasks, connect dependencies and add human approval gates. Saving a workflow stores a local reusable definition; creating a draft does not execute agents.</p><div class="columns"><div class="inspector"><label class="metric-label" for="workflowName">Workflow name</label><input id="workflowName" maxlength="100" placeholder="e.g. Secure feature delivery" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="workflowDescription">Description</label><input id="workflowDescription" maxlength="500" placeholder="What this pipeline is for" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><div class="divider"></div><h3 style="margin:4px 0">Add specialist task</h3><label class="metric-label" for="workflowAgent">Agent</label><select id="workflowAgent" class="btn" style="text-align:left"><option value="planner">Planner</option><option value="architect">Architect</option><option value="code-reviewer">Code reviewer</option><option value="security-reviewer">Security reviewer</option><option value="test-engineer">Test engineer</option><option value="documentation-writer">Documentation writer</option><option value="performance-engineer">Performance engineer</option><option value="debugger">Debugger</option><option value="devops-engineer">DevOps engineer</option><option value="database-reviewer">Database reviewer</option><option value="accessibility-reviewer">Accessibility reviewer</option><option value="release-manager">Release manager</option></select><label class="metric-label" for="workflowTaskTitle">Task title</label><input id="workflowTaskTitle" maxlength="120" placeholder="Review API boundaries" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="workflowTaskPrompt">Task instructions</label><textarea id="workflowTaskPrompt" maxlength="4000" rows="3" placeholder="Describe the specialist's job and expected evidence…" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"></textarea><label class="metric-label" for="workflowDependency">Depends on</label><select id="workflowDependency" class="btn" multiple size="4" style="text-align:left;min-height:100px"><option value="">No dependency (start independently)</option></select><div class="run-meta">Use Ctrl/Cmd to select multiple prerequisites for branching and join points.</div><label style="display:flex;gap:8px;align-items:center"><input id="workflowRequiresApproval" type="checkbox"> Require human approval before this task</label><div class="actions"><button class="btn" id="addWorkflowTask">Add task</button><button class="btn" id="clearWorkflowTasks">Clear tasks</button></div><div class="actions"><button class="btn" id="templateFeature">Load feature-delivery template</button><button class="btn" id="templateSecurity">Load security-audit template</button></div></div><div class="inspector"><div class="section-head"><h3 style="margin:0">Pipeline canvas</h3><span class="pill" id="workflowTaskCount">0 TASKS</span></div><div id="workflowCanvas" class="activity"><div class="empty">Add a specialist task to start composing your pipeline.</div></div><div class="actions"><button class="btn" id="applyWorkflowLinks">Apply selected links</button></div><div class="actions"><button class="btn" id="saveWorkflow">Save workflow</button><button class="btn" id="createWorkflowRun">Create draft run</button></div><div id="workflowMessage" class="run-meta" aria-live="polite"></div><div class="section-head" style="margin-top:18px"><h3 style="margin:0">Saved workflows</h3><button class="btn" id="refreshWorkflows">Refresh</button></div><div id="savedWorkflows" class="activity"><div class="loading">Loading reusable workflows…</div></div></div></div></section>\n<section class="section card"><div class="section-head"><div><h2>Evaluation history</h2><div class="run-meta">Saved deterministic rubric results · use <code>
|
|
75
|
+
<section class="section card" id="agentConfigStudio"><div class="section-head"><h2>Agent Configuration Studio</h2><span class="pill">WORKSPACE-OWNED MARKDOWN</span></div><p class="run-meta">Edit custom agent and skill instructions in .anvil/agents and .anvil/skills (or .forge/). Built-in agents stay read-only. Saving changes local workspace files only; it does not execute agents or automatically promote a configuration.</p><div class="columns"><div class="inspector"><label class="metric-label" for="configCategory">Configuration type</label><select id="configCategory" class="btn"><option value="agents">Custom agent</option><option value="skills">Skill</option></select><label class="metric-label" for="configFile">Existing configuration</label><select id="configFile" class="btn"><option value="">Create a new file…</option></select><label class="metric-label" for="configId">File ID</label><input id="configId" maxlength="64" placeholder="e.g. api-specialist" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="configContent">Markdown instructions</label><textarea id="configContent" rows="12" maxlength="131072" placeholder="# Agent name Describe responsibilities, constraints and expected evidence…" style="width:100%;background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"></textarea><div class="actions"><button class="btn" id="newConfigFile">New file</button><button class="btn" id="saveConfigFile">Save configuration</button><button class="btn" id="refreshConfigFiles">Refresh list</button><button class="btn" id="captureConfigSnapshot">Capture snapshot</button></div><div id="configMessage" class="run-meta" aria-live="polite">Changes are saved locally and remain reviewable in Git.</div></div><div><h3 style="margin-top:0">Configuration inventory</h3><div id="configInventory" class="list"><div class="loading">Loading custom agent and skill files…</div></div><h3>Provider readiness</h3><div id="providerReadiness" class="list"><div class="loading">Checking local provider configuration…</div></div><p class="run-meta">Shows credential presence and configured models only. It does not test provider connectivity or expose API keys.</p><h3>Benchmark-gated promotion</h3><div class="inspector"><label class="metric-label" for="releaseCandidate">Candidate snapshot ID</label><input id="releaseCandidate" placeholder="cfg-…" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="releaseBaseline">Approved baseline snapshot ID</label><input id="releaseBaseline" placeholder="cfg-…" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="releaseMatrix">Benchmark matrix ID</label><input id="releaseMatrix" placeholder="matrix-id" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><button class="btn" id="prepareConfigRelease">Check regression gate & prepare release</button><div id="releaseMessage" class="run-meta" aria-live="polite">Only a release-ready benchmark with matching snapshot fingerprint can proceed.</div></div><div id="configReleases" class="list"><div class="loading">Loading release history…</div></div><h3>Version history</h3><div id="configVersions" class="list"><div class="loading">Loading snapshots…</div></div><div class="workflow-analysis"><b>Safety boundary</b><br>Only Markdown files with validated IDs under .anvil/agents (.forge/) or .anvil/skills (.forge/) can be saved. This editor cannot write arbitrary paths, run shell commands, or modify built-in agent definitions.</div></div></div></section>
|
|
76
|
+
<section class="section card" id="workflowEditor"><div class="section-head"><h2>Visual workflow editor</h2><span class="pill">REUSABLE AGENT PIPELINES</span></div><p class="run-meta">Compose specialist tasks, connect dependencies and add human approval gates. Saving a workflow stores a local reusable definition; creating a draft does not execute agents.</p><div class="columns"><div class="inspector"><label class="metric-label" for="workflowName">Workflow name</label><input id="workflowName" maxlength="100" placeholder="e.g. Secure feature delivery" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="workflowDescription">Description</label><input id="workflowDescription" maxlength="500" placeholder="What this pipeline is for" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><div class="divider"></div><h3 style="margin:4px 0">Add specialist task</h3><label class="metric-label" for="workflowAgent">Agent</label><select id="workflowAgent" class="btn" style="text-align:left"><option value="planner">Planner</option><option value="architect">Architect</option><option value="code-reviewer">Code reviewer</option><option value="security-reviewer">Security reviewer</option><option value="test-engineer">Test engineer</option><option value="documentation-writer">Documentation writer</option><option value="performance-engineer">Performance engineer</option><option value="debugger">Debugger</option><option value="devops-engineer">DevOps engineer</option><option value="database-reviewer">Database reviewer</option><option value="accessibility-reviewer">Accessibility reviewer</option><option value="release-manager">Release manager</option></select><label class="metric-label" for="workflowTaskTitle">Task title</label><input id="workflowTaskTitle" maxlength="120" placeholder="Review API boundaries" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><label class="metric-label" for="workflowTaskPrompt">Task instructions</label><textarea id="workflowTaskPrompt" maxlength="4000" rows="3" placeholder="Describe the specialist's job and expected evidence…" style="background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"></textarea><label class="metric-label" for="workflowDependency">Depends on</label><select id="workflowDependency" class="btn" multiple size="4" style="text-align:left;min-height:100px"><option value="">No dependency (start independently)</option></select><div class="run-meta">Use Ctrl/Cmd to select multiple prerequisites for branching and join points.</div><label style="display:flex;gap:8px;align-items:center"><input id="workflowRequiresApproval" type="checkbox"> Require human approval before this task</label><div class="actions"><button class="btn" id="addWorkflowTask">Add task</button><button class="btn" id="clearWorkflowTasks">Clear tasks</button></div><div class="actions"><button class="btn" id="templateFeature">Load feature-delivery template</button><button class="btn" id="templateSecurity">Load security-audit template</button></div></div><div class="inspector"><div class="section-head"><h3 style="margin:0">Pipeline canvas</h3><span class="pill" id="workflowTaskCount">0 TASKS</span></div><div id="workflowCanvas" class="activity"><div class="empty">Add a specialist task to start composing your pipeline.</div></div><div class="actions"><button class="btn" id="applyWorkflowLinks">Apply selected links</button></div><div class="actions"><button class="btn" id="saveWorkflow">Save workflow</button><button class="btn" id="createWorkflowRun">Create draft run</button></div><div id="workflowMessage" class="run-meta" aria-live="polite"></div><div class="section-head" style="margin-top:18px"><h3 style="margin:0">Saved workflows</h3><button class="btn" id="refreshWorkflows">Refresh</button></div><div id="savedWorkflows" class="activity"><div class="loading">Loading reusable workflows…</div></div></div></div></section>\n<section class="section card"><div class="section-head"><div><h2>Evaluation history</h2><div class="run-meta">Saved deterministic rubric results · use <code>anvil evaluate RUNID</code> to score a run</div></div><span id="evaluationUpdated" class="pill">LOCAL REPORTS</span></div><div id="evaluationHistory" class="activity"><div class="loading">Loading evaluation reports…</div></div></section>\n<section class="section card"><div class="section-head"><div><h2>Benchmark Lab</h2><div class="run-meta">Compare recorded runs by provider and model · use <code>anvil benchmark RUN_ID</code> after evaluation</div></div><span id="benchmarkUpdated" class="pill">NO TRIALS</span></div><div id="benchmarkHistory" class="activity"><div class="loading">Loading benchmark history…</div></div></section>
|
|
77
77
|
<section class="section card"><div class="section-head"><div><h2>Regression intelligence</h2><div class="run-meta">Compare the latest matrix with a previous run using the same goal and evaluation suite</div></div><span id="regressionStatus" class="pill">CHECKING HISTORY</span></div><div id="benchmarkRegression" class="activity"><div class="loading">Analyzing benchmark history…</div></div><div class="actions"><button class="btn" id="refreshRegression">Refresh regression analysis</button></div></section>
|
|
78
78
|
<section class="section card"><div class="section-head"><div><h2>Agent Performance</h2><div class="run-meta">Latest matrix · deterministic evaluation scorecards by agent and model</div></div><span id="agentPerformanceUpdated" class="pill">NO SCORECARDS</span></div><div id="agentPerformance" class="activity"><div class="loading">Loading agent scorecards…</div></div></section>
|
|
79
79
|
<section class="section card"><div class="section-head"><div><h2>Skill Experiments</h2><div class="run-meta">Controlled same-model comparison · baseline with no skills versus treatment with selected skills</div></div><span id="skillExperimentsUpdated" class="pill">NO EXPERIMENTS</span></div><div id="skillExperiments" class="activity"><div class="loading">Loading controlled skill experiments…</div></div></section>\n<section class="section card"><div class="section-head"><div><h2>Experiment History & Leaderboard</h2><div class="run-meta">Historical candidate rankings · grouped by evaluation suite · estimates are not billing data</div></div><div class="actions"><span id="experimentHistoryUpdated" class="pill">LOADING HISTORY</span><button class="btn" id="refreshExperimentHistory">Refresh</button></div></div><div class="columns"><div><label class="metric-label" for="experimentSuiteFilter">Evaluation suite</label><input id="experimentSuiteFilter" placeholder="All suites" style="width:100%;background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:9px;font:inherit"></div><div><label class="metric-label" for="experimentCandidateFilter">Candidate</label><input id="experimentCandidateFilter" placeholder="All candidates" style="width:100%;background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:9px;font:inherit"></div></div><div id="experimentLeaderboard" class="activity" style="margin-top:14px"><div class="loading">Aggregating saved benchmark matrices…</div></div><div class="section" style="margin-top:18px"><div class="section-head"><div><h3>Visual candidate comparison</h3><div class="run-meta">Score and pass rate are normalized to 100 · duration and token values are estimates</div></div><span class="pill">TOP 8</span></div><div id="experimentComparison" class="activity"><div class="loading">Building comparison from saved benchmark history…</div></div></div><div id="experimentRegressions" class="activity" style="margin-top:12px"></div><div class="actions" style="margin-top:12px"><button class="btn" id="exportExperimentHistory">Export JSON</button></div><div id="experimentHistoryCaveats" class="run-meta" style="margin-top:10px"></div></section>
|
|
@@ -88,7 +88,7 @@ const PAGE = String.raw `<!doctype html>
|
|
|
88
88
|
<section class="section card"><div class="section-head"><h2>Engineering risk intelligence</h2><div class="actions"><span class="pill">STRUCTURAL HOTSPOTS</span><button class="btn" id="refreshRisk">Refresh risk analysis</button></div></div><div id="riskSummary" class="grid"><div class="loading">Calculating engineering hotspots…</div></div><div class="graph-wrap" style="margin-top:12px"><div id="riskHotspots" class="activity"><div class="loading">Ranking modules by structural risk…</div></div></div><div id="riskMethodology" class="run-meta"></div><div id="riskWarnings" class="run-meta"></div></section>
|
|
89
89
|
<section class="section card"><div class="section-head"><h2>Dependency risk explorer</h2><div class="actions"><span class="pill">INTERACTIVE PACKAGE GRAPH</span><button class="btn" id="refreshDependencyGraph">Refresh graph</button></div></div><div id="dependencySummary" class="grid"><div class="loading">Mapping package relationships…</div></div><div class="actions" style="margin:12px 0"><input id="dependencyFilter" type="search" placeholder="Filter files or packages…" style="flex:1;min-width:0;background:#0b1018;color:var(--text);border:1px solid var(--line);border-radius:9px;padding:10px;font:inherit"><select id="dependencyMode" class="btn" aria-label="Graph filter"><option value="all">All relationships</option><option value="risk">Risk signals</option><option value="packages">Packages only</option></select></div><div class="graph-wrap"><svg id="dependencyGraph" viewBox="0 0 1100 620" role="img" aria-label="Interactive file and package dependency graph"><text x="24" y="35" fill="#8e9aab">Loading dependency graph…</text></svg><div class="graph-hint">Files appear on the left and packages on the right. Select a node to inspect its consumers and status.</div></div><div id="dependencyDetail" class="graph-detail">Select a package or source module to inspect its relationships.</div><div id="dependencyWarnings" class="run-meta"></div></section>
|
|
90
90
|
<section class="section card"><div class="section-head"><h2>Architecture snapshot</h2><div class="actions"><span class="pill">STATIC IMPORT MAP</span><button class="btn" id="refreshMap">Refresh map</button></div></div><div class="graph-wrap"><svg id="architectureGraph" viewBox="0 0 900 360" role="img" aria-label="Interactive source module dependency graph"><text x="24" y="35" fill="#8e9aab">Loading architecture graph…</text></svg><div class="graph-hint">Select a node to inspect imports and incoming dependencies. Lines represent detected relative imports.</div></div><div id="graphDetail" class="graph-detail">Choose a module in the graph.</div><div id="architecture" class="architecture"></div></section>
|
|
91
|
-
<footer class="footer"><span>ANVIL
|
|
91
|
+
<footer class="footer"><span>ANVIL v1.0.0 · ENGINEERING INTELLIGENCE</span><span>Read-only dashboard · Data stays on this machine</span></footer>
|
|
92
92
|
</main>
|
|
93
93
|
<script>
|
|
94
94
|
const el=id=>document.getElementById(id);
|
|
@@ -96,8 +96,8 @@ const esc=value=>String(value??"").replace(/[&<>"']/g,c=>({"&":"&","<":"<
|
|
|
96
96
|
async function api(path){const response=await fetch(path,{headers:{"accept":"application/json"}});if(!response.ok)throw new Error("Request failed ("+response.status+")");return response.json()}
|
|
97
97
|
function connectSSE(){const es=new EventSource('/api/events');const pill=el("liveStatus");const setLive=at=>{pill.className="status live-pulse";pill.style.color="#4ade80";pill.style.borderColor="#2d6a1f";pill.textContent="● LIVE · "+new Date(at).toLocaleTimeString();};es.addEventListener("connected",e=>{try{const d=JSON.parse(e.data);setLive(d.at);}catch{}});es.addEventListener("run-updated",e=>{try{const d=JSON.parse(e.data);setLive(d.at||new Date().toISOString());loadRuns();loadActivity();if(selectedRunId)inspectRun(selectedRunId);}catch{}});es.addEventListener("heartbeat",e=>{try{const d=JSON.parse(e.data);setLive(d.at);}catch{}});es.onerror=()=>{pill.className="status";pill.style.color="";pill.style.borderColor="";pill.textContent="● LOCAL WORKSPACE";setTimeout(connectSSE,5000);es.close();};}
|
|
98
98
|
function metric(label,value,sub){return '<div class="card"><div class="metric-label">'+esc(label)+'</div><div class="metric">'+esc(value)+'</div><div class="metric-sub">'+esc(sub)+'</div></div>'}
|
|
99
|
-
async function loadEvaluationHistory(){try{const reports=await api("/api/evaluations");el("evaluationUpdated").textContent=reports.length+" saved reports";el("evaluationHistory").innerHTML=reports.length?reports.map(report=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(report.suiteName)+'</b><span class="pill">'+(report.passed?'PASS':'NEEDS REVIEW')+'</span></div><div class="run-meta">'+esc(report.runId)+' · '+esc(new Date(report.evaluatedAt).toLocaleString())+'</div><div class="run-meta">Score '+report.score+'/100 · '+report.summary.passed+'/'+report.summary.total+' deterministic cases passed · '+report.summary.missingOutput+' missing outputs</div>'+(report.judge?'<div class="run-meta">Model judge: '+esc(report.judge.provider)+' · '+report.judge.averageScore+'/100 · '+(report.judge.passed?'PASS':'NEEDS REVIEW')+'</div>':'')+report.cases.map(item=>'<div class="run-meta">'+esc(item.caseId)+' · '+esc(item.agent)+' · '+item.score+'/100 · '+(item.passed?'PASS':esc(item.status))+'</div>').join("")+'</div></div>').join(""):'<div class="empty">No evaluation reports yet. Run <code>
|
|
100
|
-
async function loadBenchmarkHistory(){try{const [docs,matrices]=await Promise.all([api("/api/benchmarks"),api("/api/benchmark-matrices")]);const groups=docs.flatMap(doc=>doc.groups.map(group=>({suite:doc.suiteName,key:group.key,trials:group.trials,completed:group.trials,averageScore:group.averageScore,passRate:group.passRate,scoreStdDev:group.scoreStdDev,bestScore:group.bestScore,worstScore:group.worstScore,averageDurationMs:group.averageDurationMs,averageEstimatedTokens:group.averageEstimatedTokens,totalRetries:group.totalRetries,averageJudgeScore:group.averageJudgeScore,judgePassRate:group.judgePassRate})));const matrixGroups=matrices.flatMap(matrix=>matrix.summary.map(group=>({suite:matrix.matrixName+" · "+matrix.suiteId,key:group.candidateId+" ("+group.provider+"/"+group.model+")",trials:group.trials,completed:group.completed,averageScore:group.averageScore,passRate:group.passRate,scoreStdDev:group.scoreStdDev,bestScore:group.averageScore,worstScore:group.averageScore,averageDurationMs:group.averageDurationMs,averageEstimatedTokens:group.averageEstimatedTokens,totalRetries:0})));const all=[...groups,...matrixGroups];const trials=docs.reduce((n,doc)=>n+doc.trials.length,0)+matrices.reduce((n,matrix)=>n+matrix.results.length,0);el("benchmarkUpdated").textContent=all.length+" model groups · "+trials+" trials";el("benchmarkHistory").innerHTML=all.length?all.map(group=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(group.key)+'</b><span class="pill">'+group.averageScore+'/100 avg</span></div><div class="run-meta">'+esc(group.suite)+' · '+group.completed+'/'+group.trials+' completed · '+group.passRate+'% pass rate · score σ '+group.scoreStdDev+(group.averageJudgeScore===undefined?'':' · judge avg '+group.averageJudgeScore+'/100 ('+group.judgePassRate+'% pass)')+'</div><div class="run-meta">Best '+group.bestScore+' · worst '+group.worstScore+' · avg '+group.averageDurationMs+' ms · ~'+group.averageEstimatedTokens+' tokens · '+group.totalRetries+' retries</div></div></div>').join(""):'<div class="empty">No benchmark trials yet. Run <code>
|
|
99
|
+
async function loadEvaluationHistory(){try{const reports=await api("/api/evaluations");el("evaluationUpdated").textContent=reports.length+" saved reports";el("evaluationHistory").innerHTML=reports.length?reports.map(report=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(report.suiteName)+'</b><span class="pill">'+(report.passed?'PASS':'NEEDS REVIEW')+'</span></div><div class="run-meta">'+esc(report.runId)+' · '+esc(new Date(report.evaluatedAt).toLocaleString())+'</div><div class="run-meta">Score '+report.score+'/100 · '+report.summary.passed+'/'+report.summary.total+' deterministic cases passed · '+report.summary.missingOutput+' missing outputs</div>'+(report.judge?'<div class="run-meta">Model judge: '+esc(report.judge.provider)+' · '+report.judge.averageScore+'/100 · '+(report.judge.passed?'PASS':'NEEDS REVIEW')+'</div>':'')+report.cases.map(item=>'<div class="run-meta">'+esc(item.caseId)+' · '+esc(item.agent)+' · '+item.score+'/100 · '+(item.passed?'PASS':esc(item.status))+'</div>').join("")+'</div></div>').join(""):'<div class="empty">No evaluation reports yet. Run <code>anvil evaluate RUNID</code> after creating a saved run.</div>';}catch(error){el("evaluationHistory").innerHTML='<div class="error">'+esc(error.message)+'</div>';}}
|
|
100
|
+
async function loadBenchmarkHistory(){try{const [docs,matrices]=await Promise.all([api("/api/benchmarks"),api("/api/benchmark-matrices")]);const groups=docs.flatMap(doc=>doc.groups.map(group=>({suite:doc.suiteName,key:group.key,trials:group.trials,completed:group.trials,averageScore:group.averageScore,passRate:group.passRate,scoreStdDev:group.scoreStdDev,bestScore:group.bestScore,worstScore:group.worstScore,averageDurationMs:group.averageDurationMs,averageEstimatedTokens:group.averageEstimatedTokens,totalRetries:group.totalRetries,averageJudgeScore:group.averageJudgeScore,judgePassRate:group.judgePassRate})));const matrixGroups=matrices.flatMap(matrix=>matrix.summary.map(group=>({suite:matrix.matrixName+" · "+matrix.suiteId,key:group.candidateId+" ("+group.provider+"/"+group.model+")",trials:group.trials,completed:group.completed,averageScore:group.averageScore,passRate:group.passRate,scoreStdDev:group.scoreStdDev,bestScore:group.averageScore,worstScore:group.averageScore,averageDurationMs:group.averageDurationMs,averageEstimatedTokens:group.averageEstimatedTokens,totalRetries:0})));const all=[...groups,...matrixGroups];const trials=docs.reduce((n,doc)=>n+doc.trials.length,0)+matrices.reduce((n,matrix)=>n+matrix.results.length,0);el("benchmarkUpdated").textContent=all.length+" model groups · "+trials+" trials";el("benchmarkHistory").innerHTML=all.length?all.map(group=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(group.key)+'</b><span class="pill">'+group.averageScore+'/100 avg</span></div><div class="run-meta">'+esc(group.suite)+' · '+group.completed+'/'+group.trials+' completed · '+group.passRate+'% pass rate · score σ '+group.scoreStdDev+(group.averageJudgeScore===undefined?'':' · judge avg '+group.averageJudgeScore+'/100 ('+group.judgePassRate+'% pass)')+'</div><div class="run-meta">Best '+group.bestScore+' · worst '+group.worstScore+' · avg '+group.averageDurationMs+' ms · ~'+group.averageEstimatedTokens+' tokens · '+group.totalRetries+' retries</div></div></div>').join(""):'<div class="empty">No benchmark trials yet. Run <code>anvil benchmark-matrix</code> or record a completed run with <code>anvil benchmark RUN_ID</code>.</div>';}catch(error){el("benchmarkHistory").innerHTML='<div class="error">'+esc(error.message)+'</div>';}}
|
|
101
101
|
async function loadBenchmarkRegression(){try{const report=await api("/api/benchmark-regression");const status=report.status.toUpperCase().replaceAll("-"," ");el("regressionStatus").textContent=status;el("benchmarkRegression").innerHTML='<div class="activity-item"><div><b>'+esc(report.matrixName)+'</b><div class="run-meta">Suite '+esc(report.suiteId||"—")+' · baseline '+(report.baselineAt?esc(new Date(report.baselineAt).toLocaleString()):"not available")+' · current '+(report.currentAt?esc(new Date(report.currentAt).toLocaleString()):"not available")+'</div>'+report.summary.map(item=>'<div class="run-meta">'+esc(item)+'</div>').join("")+'</div></div>'+(report.candidates.length?report.candidates.map(item=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(item.candidateId)+' · '+esc(item.provider)+'/'+esc(item.model)+'</b><span class="pill">'+(item.scoreDelta>0?"+":"")+item.scoreDelta+' score</span></div><div class="run-meta">Pass rate '+item.baselinePassRate+'% → '+item.currentPassRate+'% · duration '+item.durationChangePercent+'% · estimated tokens '+item.tokenChangePercent+'%</div>'+item.regressions.map(issue=>'<div class="run-meta" style="color:var(--danger)">REGRESSION · '+esc(issue)+'</div>').join("")+item.warnings.map(issue=>'<div class="run-meta" style="color:var(--warning)">WARNING · '+esc(issue)+'</div>').join("")+'</div></div>').join(""):'<div class="empty">No comparable candidates found. Run the same matrix again with the same goal and suite to establish a baseline.</div>')+(report.agents?.length?'<div class="run-meta" style="margin:14px 0 6px"><b>Per-agent scorecards</b> · skill names are observed context, not causal attribution</div>'+report.agents.map(item=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(item.candidateId)+' · '+esc(item.agent)+'</b><span class="pill">'+item.baselineScore+' → '+item.currentScore+'</span></div><div class="run-meta">Pass rate '+item.baselinePassRate+'% → '+item.currentPassRate+'% · skills observed: '+esc(item.skillsObserved.join(", ")||"none")+'</div>'+item.regressions.map(issue=>'<div class="run-meta" style="color:var(--danger)">REGRESSION · '+esc(issue)+'</div>').join("")+item.warnings.map(issue=>'<div class="run-meta" style="color:var(--warning)">WARNING · '+esc(issue)+'</div>').join("")+'</div></div>').join(""):'');}catch(error){el("benchmarkRegression").innerHTML='<div class="error">'+esc(error.message)+'</div>';}}
|
|
102
102
|
async function loadAgentPerformance(){try{const data=await api("/api/agent-performance");el("agentPerformanceUpdated").textContent=data.agents.length?data.agents.length+" scorecards":"NO SCORECARDS";el("agentPerformance").innerHTML=data.agents.length?'<div class="run-meta">'+esc(data.matrixName||"Latest matrix")+' · '+(data.createdAt?esc(new Date(data.createdAt).toLocaleString()):"")+'</div>'+data.agents.map((item,index)=>'<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>#'+(index+1)+' '+esc(item.agent)+'</b><span class="pill">'+item.averageScore+'/100</span></div><div class="run-meta">'+esc(item.candidateId)+' · '+esc(item.provider)+'/'+esc(item.model)+' · '+item.completed+'/'+item.trials+' evaluated · '+item.passRate+'% pass rate</div><div class="run-meta">Skills observed: '+esc(item.skillsObserved.join(", ")||"none")+'</div></div></div>').join(""):'<div class="empty">No per-agent scorecards yet. Run a benchmark matrix with an evaluation suite that includes agent-specific cases.</div>';}catch(error){el("agentPerformance").innerHTML='<div class="error">'+esc(error.message)+'</div>';}}
|
|
103
103
|
async function loadSkillExperiments(){try{const matrices=await api("/api/benchmark-matrices");const experiments=matrices.filter(item=>item.skillExperiment).reverse();el("skillExperimentsUpdated").textContent=experiments.length+" experiments";el("skillExperiments").innerHTML=experiments.length?experiments.map(matrix=>{const item=matrix.skillExperiment;return '<div class="activity-item"><div style="width:100%"><div class="rowitem"><b>'+esc(matrix.matrixName)+'</b><span class="pill">'+(item.scoreDelta>0?"+":"")+item.scoreDelta+' score</span></div><div class="run-meta">'+esc(matrix.suiteId)+' · '+esc(matrix.goal)+'</div><div class="run-meta">Baseline '+esc(item.baselineCandidateId)+' (no skills) → treatment '+esc(item.treatmentCandidateId)+' · skills: '+esc(item.skillIds.join(", "))+'</div><div class="run-meta">Score '+item.baselineScore+' → '+item.treatmentScore+' · pass rate '+item.baselinePassRate+'% → '+item.treatmentPassRate+'% · duration '+item.durationChangePercent+'% · estimated tokens '+item.tokenChangePercent+'%</div>'+(item.agentDeltas||[]).map(agent=>'<div class="run-meta">'+esc(agent.agent)+': '+agent.baselineScore+' → '+agent.treatmentScore+' ('+(agent.scoreDelta>0?"+":"")+agent.scoreDelta+')</div>').join("")+'<div class="run-meta">Observational benchmark only; repeat trials and inspect task outputs before attributing improvements to skills.</div></div></div>';}).join(""):'<div class="empty">No controlled skill experiments yet. Configure a skillExperiment in a benchmark matrix to compare no-skill and skill-enabled runs.</div>';}catch(error){el("skillExperiments").innerHTML='<div class="error">'+esc(error.message)+'</div>';}}
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "anvil-engineering-workbench",
|
|
3
|
-
"version": "1.0.
|
|
4
|
-
"author": "Aditya Tiwari (https://github.com/AdityaaaTiwari)",
|
|
3
|
+
"version": "1.0.2",
|
|
4
|
+
"author": "Aditya Tiwari <m.g.atiwari28@gmail.com> (https://github.com/AdityaaaTiwari)",
|
|
5
5
|
"private": false,
|
|
6
6
|
"description": "ANVIL — An open-source AI engineering harness with codebase intelligence, review history, regression signals, and a local engineering dashboard.",
|
|
7
7
|
"type": "module",
|