Skip to content

ANVIL v1.1.2

Latest

Choose a tag to compare

@github-actions github-actions released this 22 Nov 08:08

ANVIL Changelog

1.1.2 — MCP Ecosystem Hardening, Compact Tool Telemetry, Empirical Token Benchmarking & Inline Suppression Union

ANVIL 1.1.2 delivers comprehensive MCP ecosystem readiness, compact tool responses saving 90%+ model context tokens, empirical token footprint measurement, cross-platform Windows ESM fixes, and inline suppression rule unioning:

Features & MCP Improvements

  • Compact MCP Responses by Default (detail: false):
    • Core inspection tools (anvil_inspect_project, anvil_map_architecture, anvil_change_risk, anvil_git_intelligence, anvil_quality_audit, anvil_security_audit, anvil_clean_preview) return concise summaries by default, protecting AI agent context windows.
    • Every tool response includes structured telemetry in _meta: { characterCount, estimatedTokens, compact, hint }.
    • Passing detail: true returns the full AST graph, itemized findings, and file breakdowns on demand.
  • Model-Optimized Tool Descriptions:
    • Enriched tool schemas in src/mcp/server.ts with explicit model directives detailing when to call, what data is returned, and boundary guarantees (read-only, no code execution).
  • Empirical Token Footprint Benchmark (npm run bench):
    • Introduced reproducible benchmark script (scripts/measure-token-footprint.ts) measuring unassisted file loading (~220,000 tokens) vs ANVIL compact MCP output (~2,189 tokens), achieving 96–99% context token reduction across realistic engineering tasks.
  • Agent Ecosystem Assets:
    • Added root AGENTS.md (operating guide for Cursor, Claude Code, Copilot, Windsurf, Antigravity).
    • Added standard llms.txt and public/llms.txt documentation for AI search indexing.
    • Added official MCP registry manifests server.json and .well-known/mcp/server.json, plus smithery.yaml.
  • Dynamic MCP Server Versioning:
    • MCP server dynamically resolves its version from package.json (resolveVersion()), verified in CI.
  • Cross-Platform CLI & Safe Auto-Start:
    • Fixed Windows ESM path resolution in src/cli.ts by using pathToFileURL(script).href.
    • Guarded anvil mcp: displays usage and setup commands safely without hanging stdin on non-TTY calls; server runs only on explicit anvil mcp start.
  • Inline Suppression Rule Unioning:
    • Fixed regression in code-review.ts so that when both anvil-disable-next-line and anvil-disable-line apply to the same line, their suppressed rule sets and justification reasons are unioned rather than overwriting.
  • Accurate Sandbox Advisory Warnings:
    • Made sandbox resource fields (profile, maxMemory, etc.) optional in policy schema so advisory warnings are emitted only when users explicitly define resource ceilings.

1.1.1 — Governance Hardening: Disallowed Tools Matcher, SARIF inSource Suppressions & Suppression Reason Enforcement

ANVIL 1.1.1 introduces strict governance hardening, accurate disallowed tool evaluation, full SARIF in-source suppression metadata, and release consistency checks:

Features & Governance Hardening

  • Exact Command & Script Matching for sandbox.disallowedTools:
    • Evaluates package.json scripts directly for prohibited command invocations (e.g. rm, curl, wget, nc) using exact command boundary patterns.
    • Matches active finding code execution snippets using word boundary regexes instead of substring searches on human-readable diagnostic messages, eliminating false positives on words such as "format" or "information".
  • SARIF v2.1.0 In-Source Suppressions (suppressions: [{ kind: "inSource" }]):
    • Exported SARIF reports now include all inline-suppressed findings (// anvil-disable-line, etc.) tagged with standard OASIS SARIF inSource suppression metadata, preserving visibility in GitHub Code Scanning instead of vanishing silently.
    • Justification reasons documented via -- reason or : reason are captured and attached as suppressions[].justification.
  • Mandatory Justification Reasons (quality.requireSuppressionReason):
    • New policy rule quality.requireSuppressionReason: true enforces that every inline suppression comment provides an explicit justification reason (e.g., // anvil-disable-line ANVIL-R001 -- safe regex description), failing the gate if justifications are omitted.
  • Sandbox Limits Advisory Notices & Schema Clarification:
    • Dynamic container resource limits (profile, requireRootless, maxMemory, maxCpus, maxProcesses) generate clear advisory warnings in anvil gate output explaining that they govern containerized verification runs (container-sandbox) rather than static code review gates.
  • Triage Documentation Precision:
    • Updated all reproduction commands in docs/false-positive-analysis.md to reference the canonical package name npx anvil-engineering-workbench gate ..
  • Automated Action & Package Version Synchronization:
    • Added release workflow and CI assertions verifying action.yml default version remains in 100% sync with package.json and git tags.

1.1.0 — Enterprise Adoption Readiness: Policy-as-Code, Fail-Closed CI Gate & Content-Hashed Baseline Engine

ANVIL 1.1.0 delivers enterprise-grade CI/CD adoption readiness, policy-as-code enforcement, deterministic baseline suppression, and verified supply-chain hygiene:

Features & Improvements

  • Zod-Powered Policy-as-Code Engine (.anvil/policy.json & --policy):
    • Fully implemented schema validation using Zod (GatePolicySchema) enforcing strict types for security thresholds (maxCritical, maxHigh, maxMedium, failOnSecrets), code quality (maxReviewHigh, requireCleanGit), and runtime sandbox requirements (sandbox.required).
    • Strict fail-closed policy loading: malformed JSON or invalid schema in .anvil/policy.json fails the gate with explicit diagnostic messages instead of silently falling back to defaults.
    • Active enforcement of failOnSecrets (prohibiting unsuppressed credentials) and requireCleanGit (requiring pristine git status).
  • Content-Hashed Stable Baseline Engine (--baseline & --generate-baseline):
    • Finding fingerprints now compute a SHA-256 content hash of the matching code snippet (${source}:${rule}:${file}:${hash}) rather than relying strictly on volatile line numbers.
    • Moving code lines or editing adjacent blocks preserves baseline suppression without spurious regressions.
    • Backward compatible with legacy line-based fingerprints (${source}:${rule}:${file}:${line}).
  • Strict Fail-Closed CI Gate (anvil gate):
    • Code review and security scanner failures immediately fail the gate instead of falling back to empty reports.
    • If any analysis engine encounters an execution error, the gate exits with code 1 and records the failure in evaluation.violations.
  • OASIS SARIF v2.1.0 Baseline Filtering:
    • generateSarif now filters out baselined findings or marks them with standard SARIF suppressions metadata (kind: "external", status: "accepted").
    • Eliminates false-positive code scanning alerts on GitHub pull requests when findings have been explicitly baselined.
    • Dynamically resolves package version from package.json for SARIF driver metadata.
  • Inline Suppression Comments (code-review.ts):
    • Support for // anvil-disable-line <RULE>, // anvil-disable-next-line <RULE>, and /* anvil-disable-file <RULE> */ across all code review checks.
  • GitHub Action Hardening (action.yml):
    • Updated default version to v1.1.0.
    • Added support for version: local (npm link) for CI testing of workspace builds.
    • Automatic fallback detection for .anvil/baseline.json and .anvil/policy.json in target repository.
    • Robust multi-platform binary lookup across Windows, macOS, and Linux runners.
  • Coverage & Supply Chain Polish:
    • Configured @vitest/coverage-v8 with enforcement thresholds (60% lines/statements/functions, 50% branches).
    • Added .git-blame-ignore-revs ignoring repository-wide formatting commits.
    • Linked lint script to run both typechecking and Prettier verification.

1.0.9 — Release Pipeline Resilience & Idempotent npm Publishing

ANVIL 1.0.9 brings release pipeline resilience and fixes CI action resolution:

  • Idempotent npm Publishing: Added automated verification check in .github/workflows/release.yml preventing duplicate publish failures if an existing version is already present on the npm registry.
  • Workflow Action Pinning Fix: Resolved exact immutable commit SHAs for CodeQL Analysis and GitHub Pages deployment actions.
  • Documentation & UI Polish: Updated live documentation and landing page assets to reflect v1.0.9 release status.

1.0.8 — Adoption Readiness: CI Gate, OASIS SARIF v2.1.0, String-Safe Cleaner & FORGE Deprecation

ANVIL 1.0.8 focuses on enterprise company-adoption readiness, addressing third-party audit findings across supply-chain security, automated CI/CD gating, string-safe code cleanup, and complete transition away from legacy FORGE prefixes.

Features & Capabilities

  • CI Gate & Policy-as-Code (anvil gate & GitHub Action):
    • Added native anvil gate [path] [--sarif PATH] [--policy PATH] [--fail-on-warnings] [--json] CI verification mode with standard Unix exit codes (exit 0 on clean verification, exit 1 on blocking/high-severity findings).
    • Built-in OASIS SARIF v2.1.0 report generator (--sarif results.sarif) for direct integration with GitHub Advanced Security code scanning tabs.
    • Policy-as-code enforcement via docs/anvil-policy.schema.json and .anvil/policy.json (max critical secrets: 0, max high findings, required check passes).
    • Updated action.yml to default to operation: gate with sarif-file output and pinned commit SHAs.
  • String-Safe Code Cleaner:
    • Enhanced cleaner algorithm (src/core/project-cleaner.ts) with multiline string awareness (getProtectedMultilineLines).
    • Protects JavaScript / TypeScript template literals (`...`), Python triple quotes ("""...""" / '''...'''), and Go backticks from accidental whitespace stripping or blank-line collapsing.
  • Finished FORGE Renaming & Deprecation:
    • Modernized environment variable resolution prioritizing ANVIL_* prefixes with graceful backward-compatible fallback to FORGE_* (logging one-time deprecation warnings).
    • Registered dual MCP tools (anvil_* as canonical, with forge_* alias) ensuring backward compatibility for existing MCP client configurations.
  • Supply-Chain & Security Policy Hardening:
    • Pinned all GitHub Actions in CI/CD workflows (ci.yml, release.yml, codeql.yml, pages.yml) to immutable full 40-character commit SHAs.
    • Loosened CI security audit gate to npm audit --audit-level=high --omit=dev to prevent transient dev-dependency advisories from blocking production builds.
    • Added lint, format:check, and test:coverage scripts and contributor metadata to package.json.
    • Updated SECURITY.md with realistic solo maintainer SLAs (48–72h initial response) and explicitly documented GitHub Private Vulnerability Reporting enablement.
    • Clarified architectural provenance in README.md affirming 100% original, clean-room MIT-licensed code.

1.0.7 — Architecture Decision Records (ADR), Deep Security Scanner & Live Workflow DAG Runner

ANVIL 1.0.7 introduces an integrated Architecture Decision Records (ADR) engine following the MADR standard, a comprehensive repository security and secret scanner with automated .gitignore auto-fix, and a real-time interactive multi-agent workflow DAG runner in the web dashboard.

Features & Capabilities

  • Architecture Decision Records Engine (anvil adr & Web Dashboard):
    • MADR (Markdown Architectural Decision Records) 3.0 standard directory structure initialized via anvil adr init under docs/adr/.
    • Create structured decision records with status (proposed, accepted, rejected, superseded, deprecated), context, consequences, and alternatives via anvil adr new "Title".
    • Automated project memory synchronization via anvil adr sync (POST /api/adrs/sync), converting ephemeral decisions in .anvil/memory.json into permanent, version-controlled markdown ADRs.
    • Interactive visual ADR viewer and creator in the web dashboard under Architecture & Graph.
  • Deep Security & Secret Scanner (anvil security & Web Dashboard):
    • High-entropy regex and format scanning for exposed API keys: OpenAI (sk-...), Anthropic (sk-ant-...), GitHub (ghp_..., gho_..., github_pat_...), AWS access keys (AKIA...), Google API keys (AIza...), Slack tokens (xoxb-...), and private keys (BEGIN RSA/OPENSSH PRIVATE KEY).
    • Detection of dangerous shell practices (unvetted curl ... | bash in scripts) and unencrypted HTTP package dependencies.
    • Detection of exposed, unignored sensitive files (.env, .pem, .key, .pfx, .p12).
    • Automated one-click remediation with --fix (POST /api/security/fix), updating .gitignore to safely quarantine exposed files and credentials.
    • Interactive Security Audit card in the dashboard's Self-Healing & Security tab with real-time status and "Auto-Fix via .gitignore" action.
  • Interactive Live DAG Workflow Runner:
    • One-click live execution (POST /api/workflows/:id/execute) directly from the dashboard visual canvas and saved workflow cards.
    • Real-time Server-Sent Events (SSE) telemetry broadcast: workflow_started, workflow_task_settled, and workflow_completed.
    • Realistic multi-agent task execution simulation powered by advanceRun and DemoAgentProvider.
  • Interactive Landing Page Enhancements: Added terminal tabs (09 adr, 10 security) and updated live npm badges to v1.0.7.

1.0.6 — AI Harness MCP Wizard & Dashboard Project Memory Hub

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.

Features & Capabilities

  • Automated AI Harness MCP Setup (anvil mcp setup): Single-command configuration for developer environments:
    • --client cursor: Generates and non-destructively merges into .cursor/mcp.json.
    • --client claude: Generates the exact stdio command for Claude Code CLI registration.
    • --client windsurf & --client vscode: Configures .windsurf/mcp_config.json and .vscode/mcp.json.
    • --client all: Configures all supported IDE harnesses in one execution.
  • Project Memory & Architectural Decisions Hub (anvil memory & Web UI):
    • Live interactive Project Memory panel directly in the web dashboard under Overview & Health.
    • Dropdown categorization (decision, architecture, constraint, preference, fact), input tags, and single-click addition/removal.
    • Seamless dual-path storage supporting modern .anvil/memory.json with fallback to .forge/memory.json.
    • Real-time token-budgeted context ranking so agent instructions receive prioritized architectural constraints without exceeding window limits.
  • Interactive Landing Page Demos: Enhanced product site with interactive terminal simulation tabs (05 mcp-setup, 06 memory) and stage explorer showcasing harness connections.

1.0.5 — Seamless Single-Page Dashboard & Showcase Telemetry Hub

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.

Features & Improvements

  • Seamless Single-Page Dashboard: Converted the dashboard from hidden tabs into an all-in-one view where System Health, Self-Healing Cleaner, Specialist Agents & Workflow Studio, Architecture Graph, Benchmarks & Evaluation, and Workspace Explorer are all visible simultaneously, accompanied by sticky smooth-scroll navigation pills.
  • ✨ Showcase Telemetry Mode: Embedded an interactive toggle (active by default when 0 local runs exist) that simulates multi-agent execution traces, benchmark candidate matrices, token budget gauges, and self-healing verification trails so developers can explore ANVIL's full power out of the box.
  • Automatic Port Fallback Scanner: Prevented EADDRINUSE fatal startup errors by automatically scanning available ports in range 4173..4183 if the default port is held by another process.
  • Self-Healing Cleaner Dashboard Hub: Direct interactive browser controls to preview junk cleanup (/api/clean/preview), trigger safe repairs, list snapshots (/api/clean/snapshots), and execute single-click undo rollback directly from the web interface.
  • Client Script Hardening: Eliminated font string escaping anomalies and reinforced element hydration lifecycle for error-free dashboard startup.

1.0.4 — Zero-Fear Rollback, Self-Healing Verification Loop & Framework Entrypoint Engine

ANVIL 1.0.4 introduces production-grade safety nets and developer velocity superpowers: instant single-command rollback (anvil undo), an automated self-healing test verification gate (--verify), Git pre-commit hook automation (anvil hook install), and framework-aware entrypoint resolution for modern full-stack frameworks.

Features & Capabilities

  • Instant Rollback Engine (anvil undo): Every destructive action (anvil clean --apply or anvil fix) automatically creates a byte-accurate snapshot in .anvil/backups/<backup-id>/ before touching disk. Running anvil undo instantly restores deleted files, re-creates pruned directories, and reverts modified source code with 100% byte fidelity.
  • Self-Healing Test Verification Gate (--verify): When --verify is passed to anvil clean or anvil fix, ANVIL automatically detects the project's test suite (npm test, cargo test, pytest, go test ./...) and runs it after applying repairs. If tests fail, ANVIL automatically rolls back all changes before the developer even notices, guaranteeing zero regression.
  • Framework-Aware Entry Point Manifests: Solves false-positive orphan file detection across modern full-stack frameworks including Next.js, Remix, Nuxt, SvelteKit, FastAPI, Django, Celery, and Go/Rust CLI binaries. Convention-based routes (page.tsx, layout.tsx, route.ts, urls.py, conftest.py, build.rs, cmd/**/main.go) are automatically recognized as entrypoints.
  • Automated Git Pre-commit Hook (anvil hook install): Generates and installs a lightweight .git/hooks/pre-commit hook that intercepts commits and prevents OS junk, partial downloads, and syntax errors from entering Git staging.

1.0.3 — Multi-Language Architecture Mapping, Workspace Cleaner & Self-Healing Wizard

ANVIL 1.0.3 introduces comprehensive multi-language architecture mapping (Python, Go, Rust, Java, C/C++), an automated junk/clutter-file scrubber (anvil clean), automated code repair (anvil fix), and a one-command environment doctor auto-fix wizard (anvil doctor --fix).

Features & Capabilities

  • Multi-Language Architecture Mapping (anvil map): Expanded from JS/TS to Python (with relative package imports), Go, Rust (mod resolution), Java/Kotlin, and C/C++ (#include). Internal import resolution accurately maps hundreds of module relationships across mixed-language repositories with 0 false-positive circular loops.
  • Automated Cleaner Engine (anvil clean): Scans and safely prunes OS clutter (.DS_Store, Thumbs.db), incomplete downloads (.crdownload, .part, .tmp), error logs, editor swap/backup files, multi-language build caches (Python __pycache__/*.pyc, Java *.class, Rust/C++ *.o/*.obj/*.pdb), and empty directory residue. Supports safe dry-run preview by default, with --apply to execute.
  • Code Auto-Repair (anvil fix & anvil clean --fix): Safely detects and prunes unused named imports in TypeScript/JavaScript and Python (from x import a, b), preserves TypeScript type specifiers, normalizes trailing whitespace, collapses excessive blank lines, and guarantees trailing newlines without breaking compilations.
  • Self-Healing Doctor Wizard (anvil doctor --fix): Scaffolds missing .anvil/ configuration directories (agents/, skills/, workflows/), creates multi-language safe .gitignore, generates .env.example templates, and initializes Git if absent.
  • Dashboard Stability & Null-Safety: Resolved deferred element listener binding in src/core/dashboard-server.ts with universal safe listener wrappers (on()), ensuring real-time dashboard telemetry populates instantly without script halts.
  • CLI Brand & Help Unification: Aligned all commands, help strings, and examples under the official anvil command.

1.0.2 — Maintainer & Contact Verification

ANVIL 1.0.2 updates official maintainer contact information, aligning all security, support, and package metadata with Aditya Tiwari (m.g.atiwari28@gmail.com).

Changes

  • Updated maintainer and support contacts across README.md, SECURITY.md, and package.json to Aditya Tiwari (m.g.atiwari28@gmail.com).
  • Refreshed package metadata for npm registry synchronization.

1.0.1 — Rebranding, Modern Website & Documentation Polish

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.

Enhancements & Fixes

  • 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.
  • 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.
  • Brand Unification: Purged all remaining legacy branding across docs/, action.yml, workflow files, and dashboard server templates.
  • High-Resolution Banner: Generated a brand new, crisp screenshot of the ANVIL engineering dashboard (docs/anvil-dashboard.png) for repository and npm README presentations.
  • Cross-Platform Test Stabilization: Resolved Windows symlink traversal and 8.3 path canonicalization issues in CI workflows.

1.0.0 — First major release

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.

Product experience

  • Local engineering dashboard with project overview, health snapshots, architecture and dependency graphs, run history, task inspection, approvals, traces, reviews, verification history, and benchmark/evaluation views.
  • Workspace explorer and code search for inspecting project files and symbols.
  • Agent Configuration Studio for custom agent and skill Markdown, snapshots, and benchmark-gated configuration promotion.
  • Visual workflow editor for reusable specialist-agent pipelines and draft run creation.
  • Provider readiness and diagnostics views that avoid exposing credential values.

Engineering capabilities

  • Project inspection, architecture mapping, dependency intelligence, engineering-risk analysis, change intelligence, and quality auditing.
  • Specialist-agent orchestration with saved runs, incremental task settlement, approval gates, replay/retry, and resume recovery.
  • Project memory, workspace skills, declarative hooks, permission policies, and bounded read-only workspace tools.
  • Review history, dry-run patch validation, explicit patch approval records, and staged promotion APIs.
  • Verification history, evaluation suites, benchmark matrices, regression comparisons, and experiment manifests.
  • MCP server and CLI workflows for local engineering automation.

Install and launch

npm install -g anvil-engineering-workbench
anvil --version
anvil init ./my-project
anvil dashboard ./my-project

The dashboard binds to localhost by default. Open the URL printed in the terminal.

Important V1 boundaries

  • The public Vercel website is a static product site. The live dashboard runs locally against a workspace selected on the user's machine; it is not a hosted multi-user dashboard.
  • Provider-backed runs require the user's own supported provider configuration and credentials. No provider connection or API spend is included by this release.
  • Native model tool-calling is currently limited to bounded read-only workspace tools for the OpenAI-compatible provider. Other documented providers remain proposal-only in the normal run path.
  • MCP workspace tools are read-only. Patch validation is a dry run; patch promotion is a separately reviewed library API, not an automatic agent action.
  • Local script verification executes project-controlled code on the host. Docker verification is opt-in and depends on the host runtime. Do not treat local execution or heuristic risk scores as a security guarantee.
  • Rootless Docker, live provider end-to-end validation, and hosted multi-user operation are not claimed as verified V1 capabilities.

This release establishes the ANVIL product foundation. Future releases can expand integrations and hosted operation without weakening the local-first permission and approval model.