trinity-vsa-framework

Plugin trinity-vsa-framework in gHashTag/trinity-fpga. Commands read from the repo's manifest — 1 month ago.

Skills (33)

• fpga-synth

  Run FPGA synthesis pipeline from .tri spec to bitstream. Use when working with Verilog, FPGA, or hardware synthesis.

• trinity-test

  Run Trinity test suites and analyze results. Use for testing VSA, VM, Firebird, WASM, or full test suite.

• vibee-gen

  Generate Zig or Verilog code from VIBEE specifications. Use when creating or updating .tri specs and regenerating code.

• vsa-verify

  Verify VSA (Vector Symbolic Architecture) mathematical proofs and operations. Use for math verification, bind/unbind/bundle testing, phi identity checks.

• agents

  Agent swarm development dashboard — live Railway containers, issue queue, PR pipeline, JSONL events, pool utilization, and queue drain status. Use when checking agent dev tasks, spawning status, or monitoring issues #315-#319.

• arena

  Trinity Arena 2.0 — LLM Battle Platform dashboard. Live leaderboard, battle history, scientific metrics, deploy to Railway.

• board

  GitHub Project Board auto-sync — fill empty fields, sync labels with board status, audit field completeness. Keeps board as realtime source of truth.

• bridge

  Trinity Bridge manual — architecture, endpoints, commands, patterns for Perplexity-Railway-Mac-Claude Code channel. Reference for all AI agents.

• ci-gates

  How trinity-fpga's CI gates actually behave — which checks are pre-existing failures, why adding a build.zig target can break the build, and how to verify a Zig change the way CI will. Use before pushing anything that touches build.zig, .github/workflows/, or a file with a pub fn main.

• cloud

  Show Trinity Cloud Dev dashboard — containers, training farm, issues, PRs, costs. Use when checking agent swarm status, cloud containers, or CI/CD pipeline.

• doctor

  HEALER — diagnose, heal, commit dirty files, monitor junk & docs & duplicates, report honestly. Every loop = action + proof.

• farm-garden
• farm

  HSLM training farm management — service status, evolution, leaderboard, recycle, inject configs, kill underperformers. Central dashboard for 152 Railway training workers across 8 accounts.

• fpga-bufr

  Regional clock buffers (BUFR/BUFIO) on openXC7 — how to make one actually clock fabric, the two placer fixes it needs, and how to prove a candidate segbits row on silicon with a one-bit A/B. Use for BUFR/BUFIO work, prjxray segbits verification, or any "is this bit real?" question on an AX7203.

• god-mode

  GOD MODE — Agent monitoring dashboard. Shows swarm status, task queue, pipeline compliance, circuit breakers, git activity, and rule violations.

• implement-issue

  Read a GitHub issue, create a branch, implement the solution, and open a PR. Use when given an issue number to implement.

• measurement-discipline

  How to make a quantisation/format measurement that survives its own audit. Self-test gates, competitor implementation from spec, confound control, provenance search, and the failure catalogue from the 2026-08 campaign. Use before writing any harness that will produce a number someone quotes.

• negative-control

  Before trusting any check, gate, test, or measurement, prove it can fail. Six gates in this repo were green because they were structurally incapable of reporting a problem. Use whenever you write a CI gate, a test, a census script, or are about to report "N problems found" — especially when N is zer

• railway-ssh

  Execute commands on Railway cloud server via SSH. Handles auth, connection, tmux sessions. Use for deploying, updating code, managing ORACLE dashboard remotely.

• review-code

  Review recent code changes for bugs, style issues, and improvements. Use after writing code or before committing.

• run-tests

  Build the project, check for warnings, run all tests, and report results. Quick health check for the codebase.

• scholar

  Self-Evolving Research Agent — scans web for relevant tech, evaluates findings, proposes improvements. Uses Perplexity Sonar API via MCP.

• sec-monitor

  Security monitoring daemon — checks for new secrets in commits, unsafe patterns in changed files, Docker/CI drift, and policy violations. Lightweight recurring scan for /loop usage.

• security-audit

  Security vulnerability scanner — scans repo for hardcoded secrets, injection vectors, unsafe patterns, Docker/CI misconfigs, and policy violations. Use when checking security posture or before releases.

• stale-reference

  Repairing a build, tree, or dependency that rotted silently — why extraction leaves dangling references, why a gate that answers two questions hides the interesting one, and the habits that separate a real fix from one that compiles. Use when something has not been built or run in a long time, when

• status

  Live system dashboard — 12-dimension ouroboros score, verdict v3, agents, build health, recommendations

• t27-spec

  Record verified claims as .t27 SSOT specs — the project's own language. Idioms, provenance discipline, status tags, and the verification traps that produced wrong counts before.

• tree

  Technology dependency tree — shows issue DAG, prereqs, critical path, blocked/ready tasks. Single source of truth for work order.

• train

  HSLM training analytics dashboard — live process status, checkpoint loss curve, model architecture, platform benchmarks, FPGA status, and scientific metrics. Use when checking training progress, loss convergence, or preparing paper-ready results.

• tri

  TRI status dashboard — agent thought mode by default, /tri full for visual dashboard

• trinet

  TRI-NET ternary internet — node cell, receipts, mesh, TRI settlement, IGLA CODER agent. Board truth, the honest status of every claim, and the next wave. Use when working on ternary compute nodes, compute receipts, the TRI credit ledger, or growing the network.

• ux-wave

  Autonomous UX improvement waves — launch 4 parallel agents per wave, audit every 5 waves, commit fixes. For Queen UI chat & network polish.

• wave

  Farm wave management — create Railway accounts, spawn workers, deploy configs, track wave progress. Use for scaling training farm.

MCP servers (3)

• trinity

  runs: /data/trinity/zig-out/bin/trinity-mcp

• zig-docs

  runs: npx -y @nichochar/zig-mcp@latest

• railway-mcp-server

  runs: npx -y @railway/mcp-server