claude

harness-combined

Unified Claude Code harness: polyglot gate pipeline (lint/type-check/tests/security with diff-based repair loop and BM25 failure memory) + two-checkpoint SDLC workflow (problem → design critic → spec → build → post-build critic → deliver). Twenty-nine expert review panels covering languages, frontend frameworks, security (identity + cryptography), data/ML engineering, and operations. No API key required.

0 on GitHubby Bradley Smockv1.12.0MITupdated 1 month agodevelopment
/

Plugin harness-combined from marketplace bradleysmock-plugins in bradleysmock/claude-harness. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (11)

  • critique

    Apply domain-specialist panels to a chosen scope — Python idioms, HTTP/API design, UI patterns, AI/LLM security — each with a distinct mental model, not just a harder general scan. Scope is set by argument (files, globs, a git ref/range, or a scope keyword); when no argument is given the command ASK

  • debug

    Classify and explain an escalated spec/build run that exhausted repair attempts, then propose a targeted fix. TRIGGER when the user asks why a build failed, what to do about an escalated run, why repair gave up, how to recover from /build failure, or how to interpret an artifact in .harness/results/

  • health

    Render a read-only harness health dashboard — cross-ticket gate pass rates, average repair cycles, top recurring failure modes, tickets with the most gate failures, and per-gate trend indicators (improving / declining / stable). TRIGGER when the user asks "how healthy is the harness", "show build he

  • requirements-review

    Validate a ticket's requirements integrity before /build — a targeted checkpoint between /problem and /build that interrogates requirements.md across four named dimensions (completeness, testability, coverage, consistency) against problem.md and writes an advisory findings report. TRIGGER when the u

  • review

    Interactive panel-aware review of a ticket's implementation against its problem / requirements / solution baseline. TRIGGER when the user asks to review a ticket, check a worktree's diff before /deliver, assess whether a ticket is merge-ready, or invoke an interactive post-build review on a specific

  • rollback

    Roll back a delivered ticket by reverting its merge commit with a standardized `git revert`. TRIGGER when the user asks to "roll back ticket XXXX", "revert the delivery of XXXX", "undo a delivered ticket", or invokes /rollback (optionally with `--dry-run` to preview). SKIP when the ticket has not be

  • sprint

    Group open tickets into a time-boxed, dependency-ordered sprint plan. TRIGGER when the user asks to "plan a sprint", "build a sprint plan", "sequence the backlog", "how should we schedule the open tickets", or invokes /sprint (optionally with `--sprint-capacity N`, `--max-sprints N`, or `--as-of YYY

  • stale

    List harness tickets that have gone idle — tickets whose `updated:` date exceeds a staleness threshold (default 7 days). TRIGGER when the user asks "what's gone stale", "which tickets are idle", "show stale tickets", "any abandoned tickets", or invokes /stale (optionally with `--days N`). SKIP when

  • status

    Show the combined state of harness work — open tickets with their statuses plus recent standalone spec/build runs and failure-memory summary. TRIGGER when the user asks "what's open", "what's in flight", "what's the status", "show me tickets", "where are we?", or any general progress check across th

  • suggest

    Surface targeted, non-duplicate improvement suggestions for the current project. Detects whether the working directory is a harness plugin root or an app project, inventories accordingly, and emits /problem-ready lines for accepted suggestions. TRIGGER when the user asks "what should we build next?"

  • usage-report

    Analyze the user's own Claude Code usage across all local sessions and write a dated markdown report — usage patterns, an active-time estimate that excludes idle/away sessions, strengths/weaknesses (token efficiency + output quality), and forward-looking recommendations tied to Anthropic's public ro

Slash commands (31)

  • /harness-combined:abandon
  • /harness-combined:autopilot-watch
  • /harness-combined:autopilot
  • /harness-combined:bisect
  • /harness-combined:build
  • /harness-combined:cancel
  • /harness-combined:changelog
  • /harness-combined:deliver
  • /harness-combined:doctor
  • /harness-combined:export
  • /harness-combined:flaky
  • /harness-combined:gate
  • /harness-combined:harvest-learnings
  • /harness-combined:health
  • /harness-combined:init
  • /harness-combined:milestone
  • /harness-combined:problem
  • /harness-combined:refine
  • /harness-combined:reopen
  • /harness-combined:replan
  • /harness-combined:requirements-review
  • /harness-combined:requirements
  • /harness-combined:rollback
  • /harness-combined:solution
  • /harness-combined:sprint
  • /harness-combined:suggest
  • /harness-combined:ticket-list
  • /harness-combined:ticket-status
  • /harness-combined:usage-report
  • /harness-combined:velocity
  • /harness-combined:write-spec

Subagents (3)

  • craft

    Read-only-reasoning craft reviewer. Runs after functional acceptance (critic BLOCKER/MAJOR cleared) and proposes bounded, behaviour-preserving craft improvements — naming, structure, restraint, load-bearing comments — as structured JSON, not prose. Behaviour preservation is enforced mechanically by

  • critic

    Read-only senior-engineer reviewer. Loads expert panels by file scope, reads gate-findings if present, and produces structured BLOCKER / MAJOR / MINOR / OBS findings with file:line references. Use for both pre-implementation (design review of solution.md) and post-implementation (code review of work

  • requirements-analyst

    Read-only requirements analyst. Given a ticket's problem.md and requirements.md paths, evaluates requirements integrity across completeness, testability, coverage, and consistency and returns structured findings. Read-only — no write capability. Used by the requirements-review skill to contain the p

MCP servers (1)

  • harness

    runs: bash ${CLAUDE_PLUGIN_ROOT}/bin/harness-server

Hooks (5)

  • PreToolUse

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/pre_ticket_diff.py"

  • PreToolUse

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/pre_write_guard.py"

  • PostToolUse

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/post_write_gate.py"

  • Stop

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/stop_full_gate.py"

  • Stop

    python3 "${CLAUDE_PLUGIN_ROOT}/hooks/ticket_commit_guard.py"

View source on GitHub

More from bradleysmock/claude-harness

All 5 plugins →

Retrofit a high-integrity test suite onto an untested codebase: seam-aware, risk-ranked, oracle-grounded, and mutation-gated. Enforces Generator/Verifier role separation and the net-immutability and oracle-independence invariants mechanically.

12 commands · 2 agents · hooks

0

Structured two-checkpoint SDLC harness: slash commands, polyglot validation hooks, expert review panels, and a working agreement.

12 commands · 1 agent · hooks

0

harness

claude

MCP server wrapping the harness-full LLM coding harness: spec generation, polyglot gates, failure memory, and multi-spec task orchestration.

8 commands · 1 MCP server

0

Harness without an API key: Claude Code is the orchestrator, Python provides gates, memory, and DAG tools via MCP. Requires only a Claude Code subscription.

6 commands · 1 MCP server

0

Similar plugins

superpowers

claudecodex

Core skills library for Claude Code: TDD, debugging, collaboration patterns, and proven techniques

15 skills · hooks

obra/superpowers 316,836

Taste Skill

claudecodex

tasteskill 2.5.1: professional company websites at the figma.com and x.ai level, with one section per screen, real photographs, real product UI, effects kept for tech and gaming pages, and a real-browser checker and finish gate.

1 skill

Leonxlnx/taste-skill 114,391

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.

5 commands

HKUDS/CLI-Anything 58,665

5 skills + 35 modes + Material Passport pipeline. Includes v3.6.7 cross-model audit gate and v3.6.8 generator-evaluator contract.

5 skills · 16 commands · 4 agents · hooks

Imbad0202/academic-research-skills 56,698

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Maintain harness-combined?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex plugins5 plugins