fixing-harness

A review and refactoring tool for an AI coding project's instruction system. It works with files such as CLAUDE.md, AGENTS.md, rule files, and settings that guide coding agents.

In plain words
What is it for?
Use it to audit and reorganize an agent harness, propose a refactor, update its instruction and settings files after approval, and verify the configuration with the project's memory command.
Why use it?
It reduces the amount of irrelevant guidance an agent loads by keeping critical rules always available and assigning other rules to matching file paths. It also adds explicit permission safeguards and checks the result.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/agentic-dev3o/devx-plugins/harness-fixer
Any agent
npx skills add agentic-dev3o/devx-plugins --skill harness-fixer
Clone the repo
git clone --depth 1 https://github.com/agentic-dev3o/devx-plugins

Made for: Claude Code, Codex.

Per session 133 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,688 The whole file, excluding the scripts and references it only reads on demand.
Security scan B 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00133 $0.01688
Opus 5 $0.00067 $0.00844
Sonnet 5 $0.00027 $0.00338
Haiku 4.5 $0.00013 $0.00169

Measured 2d ago against content hash 80504aa004b0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

fixing-harness scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

3. **Deny rules in place** — `grep -c '"deny"' .claude/settings.json` should return ≥ 1. The deny list must include destructive ops and secret reads.
plugins/qa/skills/harness-fixer/SKILL.md · 133 lines

How it starts

The opening of the file, as written. The whole thing — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Fixing the Agentic Harness

Target: $ARGUMENTS (optional path; defaults to project root)

Objective

Refactor the project's harness so the model sees only the rules it needs for the files it touches. CLAUDE.md (and AGENTS.md when present) become a thin always-on critical-rules layer. Domain rules move to .claude/rules/*.md with paths: frontmatter. Permissions and MCP scope live in .claude/settings.json with explicit deny guardrails.

The goal is inference weight, not exhaustive documentation: 10 sharp directives the model references in reasoning beat 200 lines it half-skims.

Workflow

Progress checklist:

Harness Audit & Fix:
- [ ] Step 1: Inventory current harness
- [ ] Step 2: Audit against best practices
- [ ] Step 3: Produce gap analysis
- [ ] Step 4: Propose refactor plan (confirm with user)
- [ ] Step 5: Apply changes (CLAUDE.md, AGENTS.md, rules, settings)
- [ ] Step 6: Verify with /memory

Step 1: Inventory Current Harness

Read in parallel where present:

  • CLAUDE.md, ./.claude/CLAUDE.md, CLAUDE.local.md
  • AGENTS.md (Codex / Cursor / other agents)
  • .claude/settings.json, .claude/settings.local.json
  • .claude/rules/**/*.md — note paths: frontmatter on each
  • .claude/agents/*.md, .claude/skills/*/SKILL.md
  • .mcp.json, .gitignore

Capture the actual directory layout (top-level apps, packages, services). Use tokei if available, otherwise find or ls. The layout drives the path scopes proposed in Step 4 — generic globs that don't match real directories are dead weight.

Step 2: Audit Against Best Practices

Load references/audit-checklist.md and score each dimension:

  1. CLAUDE.md size and shape — under 30 lines, directives not prose, no domain dump
  2. AGENTS.md alignment — mirrors CLAUDE.md when present, no drift
  3. Rules segmentation — path-scoped, one domain per file, no overlap
  4. Permissions hygiene — explicit deny for destructive ops and secrets, scoped allow patterns, MCP tools narrowed
  5. Local overrides.claude/settings.local.json and CLAUDE.local.md in .gitignore
  6. Memory ergonomics — no monolithic root CLAUDE.md drowning monorepo subtrees

Read the full file on GitHub · 133 lines

Files

What ships with it

5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 133 lines · 133 tokens per session scan B 80504aa004b0

Subscribe to this mod's changes

fixing-harness is a skill published in the GitHub repository agentic-dev3o/devx-plugins (11 stars, last pushed 14d ago), licensed MIT. It adds 133 tokens to every session and 1,688 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

systematic-debugging

Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.

obra/superpowers · 21 tokens

next-cache-components-adoption

Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…

vercel/next.js · 95 tokens

babysit-pr

Babysit a GitHub pull request after creation by continuously polling review comments, CI checks/workflow runs, and mergeability state until the PR is merged/closed or user help is required. Diagnose failures, retry likely flaky failures up to 3 times, auto-fix/push branch-related issues when appropriate, and keep…

openai/codex · 114 tokens

imagegen

Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output…

openai/codex · 113 tokens

cpu-profile-analysis

Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…

microsoft/vscode · 71 tokens

next-cache-components-optimizer

Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…

vercel/next.js · 170 tokens