Repository navigation
Releases
Releases · openasocket/NeckBeard
NeckBeard v0.3.1 — Claude.app Dashboard Launcher Fix & Multi-Harness Auto-Updater
What's New in NeckBeard v0.3.1
1. Fixed Claude.app Banner Dashboard Launcher
- Fixed an unescaped backtick pair in
hooks/band/dashboard.tsthat threw a runtimeTypeErrorwhen clickingDashboardin the Claude Desktop / CLIAbovePromptHUD. - Hardened
launchDashboardInBrowser($)inhooks/band/register.tsxwith a fault-toleranttry/catchwrapper and a direct/usr/bin/openfallback on macOS GUI sessions. - Added automated runtime execution tests for
hooks/band/analyzer.tsandhooks/band/dashboard.tsintest/hooks.test.js.
2. Cross-Platform Multi-Harness Auto-Discovery Installer & Updater (scripts/install-all.js)
- One command (
node scripts/install-all.jsornpm run sync) runs across Windows, macOS, and Linux to:- Automatically pull the latest
main& tags from GitHub when run from a clean checkout, - Discover all installed compatible agent harnesses (
Claude Code,OpenAI Codex,GitHub Copilot CLI,Grok CLI,Antigravity CLI,Gemini CLI,Pi / Oh My Pi,OpenCode / Kilo Code), - Update their CLI plugins and local versioned plugin caches to
v0.3.1(and disable legacyponytail@ponytailcollisions), and - Rebuild the local NeckBeard Security & Engineering Command Center (
~/.config/neckbeard/telemetry/dashboard.html).
- Automatically pull the latest
NeckBeard v0.3.0 — Uncapped Audits, E_patch vs. E_rewrite Math & 5-Pillar VIBES Command Center
What's New in NeckBeard v0.3.0
1. Uncapped /neckbeard-audit (lite & complete) + Explicit E_patch vs. E_rewrite Math
- Zero Artificial Finding Cap: Removed the 20-finding cap and
Do not search past the findings in this reportrestriction fromskills/neckbeard-audit/SKILL.md. Every confirmed Must fix and Should fix finding (and all repository variants) is reported in full. When total findings exceed 25, NeckBeard states the exact total found and explicitly alerts/prompts the user to expand every remaining item in line-by-line detail or export them to a remediationSPEC.md/.maestro/playbooks/checklist. - Math-Backed Refactor vs. Rewrite Assessment: Recommending a full rewrite from scratch is exceedingly rare. Before ever suggesting one,
/neckbeard-auditcomputes and shows the explicit engineering effort/time math comparingE_patch(root fixes × affected LOC/callers + hostile-input verification tests) againstE_rewrite(clean minimal LOC to pass all 3 gates + caller migration + full test parity × regression risk).
2. 100% Original Telemetry Engine & Host Adapters
- Clean-room 100% original telemetry engine and host extensions:
hooks/band/analyzer.ts,hooks/band/dashboard.ts,hooks/band/register.tsx,hooks/band/types.d.ts,hooks/band-import.js,hooks/statusline.sh,pi-extension/index.js, and.opencode/plugins/neckbeard.mjs. - Stores local telemetry in
~/.config/neckbeard/telemetry/sessions.jsonlwith zero outbound network calls.
3. 5-Tab NeckBeard Security & Engineering Command Center (~/.config/neckbeard/telemetry/dashboard.html)
- High-Value Baseline Engineering Intelligence (Even Without VIBES): 4-card 3-Gate & Economics Scoreboard, 14-day API spend & prompt-cache efficiency SVG chart, clickable Per-Repository Engineering Scorecard, Model & Prompt-Cache Economics Breakdown Table, External
[ext]→ Internal[int]Trust-Boundary Graph, and searchable Session Forensics. - Exponential 5-Pillar Deep Forensics When VIBES (
.ai-audit/) Is Present: Full support for VIBES Protocolmanifest.entriescontent-addressed hash resolution (environment,reasoning,command,prompt,decision),line_start/line_endfile range attribution,VERIFY v1.0DSSE & high-assurance attestation coverage,PRISM v1.03-panel SVG risk histogram + turn-by-turn risk trajectory sparkline + file hotspots,EVOLVE v1.0decision ledger,TRACE v1.0incident console, and turn-level.ai-audit/annotations.jsonlprovenance inspector (plus an interactive Reference Preview & quickstart onboarding when.ai-audit/is not yet initialized).
NeckBeard v0.2.0
What's New in NeckBeard v0.2.0
1. Persistent Load Defaults & Conclusion-Biased Trust Boundaries (.neckbeard.json)
- Ask once, persist, reuse: When no load is provided in the prompt or repo and the Fast gate applies, NeckBeard asks one brief question to ascertain load, saves it to
.neckbeard.json(or via/neckbeard load <value>), and uses it as the project default across sessions and subagents. - Per-boundary load envelopes (
boundaryLoad): Override load targets per sink or boundary group via/neckbeard load <sink>=<parameters>(e.g.,/neckbeard load webhook-in=2000 rps burst). - Self-derived trust boundaries & conclusion bias: NeckBeard derives the trust boundary itself from the ask, code, and boundary map, defaulting to a fail-closed conclusion rather than pausing to ask.
2. Concise Internal & External Trust-Boundary Map, Censorship & Git Sharing Toggle
- Token-efficient two-tier map (
<= 680chars /~170tokens): Automatically maps external boundaries (ingress,egress,supply) and internal boundaries (auth,data,exec,ai) via/neckbeard map. - Automatic boundary-expansion drift alerts: Detects unreviewed new boundary sinks (
⚠ +N boundary) in prompt context, in the Claude Code and Pi banners, and during/neckbeard-review. /neckbeard-debtload-trigger expiry: Automatically comparesrisk:comment revisit thresholds against.neckbeard.jsonload and flags exceeded triggers asexpired-load-triggerunder Must fix now.- Shared by default with strict censorship:
.neckbeard.jsondefaults to shared in the repo root (ready to commit to Git) and is strictly censored byhooks/surface.jsandhooks/guard.js(repo-relative paths + closed sink enums only; URLs, hostnames, IPs, emails, absolute paths, and secrets are stripped and blocked). Toggle anytime with/neckbeard share on|off.
3. Claude Code AbovePrompt Banner, Local Telemetry Board & VIBES-First Instrumentation
- Live Claude Code banner (
hooks/claude-band.json+hooks/band/): Displays session cost ($), tokens, active agent time, definition reuse %, net refactor line delta (+added/-removed), open callers,PreToolUseguard blocks, security issues found/fixed, boundary drift, and live VIBES PRISM risk score. - 100% local HTML dashboard (
~/.config/neckbeard/band/board.html) &/neckbeard-gain: Zero network calls. Includes both a Sessions & Telemetry view and an interactive Trust-Boundary Graph SVG architecture diagram. - VIBES Protocol (
.ai-audit/) priority: Automatically reads.ai-audit/manifest.jsonand.ai-audit/annotations.jsonlwhen present for session timelines, line deltas, PRISMrisk_score, EVOLVE decisions, and TRACE incident resolution. - Native Pi / Oh My Pi (
pi-extension/) & OpenCode / Kilo Code (.opencode/) adapters.
4. /neckbeard-audit lite vs. complete Modes
lite(default): Targeted end-to-end audit of execution flows and major internal/external data & trust boundaries.complete(/neckbeard-audit complete): Exhaustive audit of every single code, configuration, script, workflow, and documentation file in the repository, with a mandatory upfront time, token, and tool-call warning.
NeckBeard v0.1.0
NeckBeard v0.1.0 — Principal-Engineer Mode for AI Coding Agents
Smallest change that is correct, safe at the boundary it crosses, fast and scalable at a named load, simple for the end user, and finished.
Built on Ponytail's 6-rung minimalist question ladder, with security, performance at a named load, and deep-module user simplicity baked in from line one.
Highlights
- Unified Portable Ruleset (
skills/neckbeard/SKILL.md&AGENTS.md):- 4 Pre-Write Framing Questions: Blast radius & user impact, trust boundary, load as a number (not 100x it), and how necessary complexity is hidden inside the module.
- 6-Rung Question Ladder:
Lean is this ladder, not a fourth gate. - Three Built-In Gates:
- Safe: Edge parsing, fail-closed errors, server-side object authorization (
Authorize the object on the server, every time. Logged in is not allowed.), parameterized SQL, shell argument arrays, path canonicalization + symlink-escape rejection, outbound host allow-lists, platform crypto, and untrusted LLM/tool outputs. - Fast and scalable at the named number: Input limits, idempotent retries, no check-then-write races, no query-per-item (N+1), no unbounded in-memory lists, no quadratic work, no blocking work on the request/UI thread, and no unindexed filters at the named load.
- Done: Fix shared code once + all broken callers/tests, single source of truth per state, backward-compatible interfaces, deep modules that hide complexity behind safe defaults, clear actionable user failure outcomes (never raw stack traces or secrets), and 1 hostile-input test on risky checks.
- Safe: Edge parsing, fail-closed errors, server-side object authorization (
- Four Focused Skills:
/neckbeard [lite|full|paranoid|off]— session coding rules and level switching (/neckbeard,$neckbeard,@neckbeard)./neckbeard-review [--ci]— change review with concrete failure proof, repo-wide variant search on Must fix findings, and--ciJSON gate output ({"verdict":"pass|fail","mustFix":N})./neckbeard-audit— whole-repo audit starting at entry points, checking the three gates, broken doc/agent-instruction promises, least-privilege permissions, supply chain, and offering an opt-inSPEC.mdor Maestro playbook remediation plan./neckbeard-debt— scans tracked and untracked files forrisk:,shortcut:, andponytail:comments and flagsillegal-deferralon trust checks or data-loss writes.
PreToolUseGuard (hooks/guard.js):full(default): Blocks hardcoded secrets (private keys, AWS, GitHub, Slack, Stripe live, Google, Anthropic, OpenAI, npm tokens) and blocks direct agent tampering with immutable VIBES Protocol.ai-audit/provenance and evidence artifacts.paranoid: Also blockscurl … | sh, disabled TLS verification,chmod 777,--no-verify, and literal credentials.
- Multi-Host Adapters: Out-of-the-box manifests for Claude Code, Codex, GitHub Copilot CLI & Editor, Antigravity CLI & IDE, Gemini CLI, Grok Build, and any
AGENTS.md-compatible host. - VIBES Protocol Integration: Zero-config support for
VIBES v1.1,VERIFY v1.0,PRISM v1.0,EVOLVE v1.0, andTRACE v1.0when.ai-audit/is present. - Zero Dependencies: Pure Node.js (
>= 18), 17 unit tests vianode:test.