harness-kit AGENTS.md

A registry and routing guide for a project whose agents, skills, and runtime paths are declared in AGENTS.md. It maps where definitions, shared files, state, and outputs live across coding tools.

In plain words
What is it for?
Use it when adding or changing agents, skills, routing pipelines, runtime state, shared documentation, or compatibility with tools such as Cursor and Codex.
Why use it?
It gives automated workflows one source of truth for choosing agents and locating their configuration and results.

Instructions file for CodexOpenCode

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 instructions/pierry/harness-kit/agents-md
Clone the repo
git clone --depth 1 https://github.com/Pierry/harness-kit

Made for: Codex, OpenCode.

Per session 3,723 This file is loaded in full into every session.
When invoked 3,723 The same file — it is already loaded in full.
Security scan C 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 $0.03723 $0.03723
Opus 5 $0.01861 $0.01861
Sonnet 5 $0.00745 $0.00745
Haiku 4.5 $0.00372 $0.00372

Measured 2d ago against content hash f688b302249f, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade C, and why

harness-kit AGENTS.md scanned grade C 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.

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

The installer pre-authorizes these in `settings.json` `permissions.allow` (both bare-path and `bash <path>` forms), so a full pipeline run never stops for a harness-internal prompt. Destructive ops (`rm -rf`, force push,
AGENTS.md · 243 lines

How it starts

The opening of the file, as written. The whole thing — 243 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md, Agent Registry & Routing

Source of truth for every agent, skill, and runtime path in this workspace.

This file is the declarative spec. Execution lives in .claude/agents/ (definitions), .claude/runtime/ (state, outputs), and .claude/shared/ (cross-agent docs). Pipelines like /sse:run and /product-manager:run consult this registry when dispatching.

Portable across tools (Claude Code, Cursor, Codex CLI, Gemini CLI, and others): all read AGENTS.md at the repo root. See Cross-tool compatibility for what each tool consumes.


Quick map

AGENTS.md                          ← this file: registry + routing
CLAUDE.md                          ← project context (style, role, conventions)
.claude/
├── agents/                        ← per-agent definitions + bundled assets
│   ├── product-manager/
│   │   ├── README.md
│   │   ├── sensors/               deterministic checks
│   │   ├── evals/                 LLM-judge rubrics
│   │   ├── guides/                pipeline, writing style, templates, examples
│   │   └── skills/                prd, prp
│   ├── staff-software-engineer/
│   │   ├── README.md
│   │   ├── sensors/
│   │   ├── evals/
│   │   ├── guides/                pipeline, coding-style, commit-style, conventions-override
│   │   └── skills/                backend, web, mobile, devops, designer (new UIs)
│   ├── system-architect/
│   │   ├── README.md
│   │   ├── sensors/               design-structure, design-rigor
│   │   ├── evals/                 design-quality, design-review-depth
│   │   ├── guides/                design-method (the canon), templates, example
│   │   └── skills/                design, review, search-engine (+ one per system design)
│   ├── product-manager.md         orchestrator agent
│   ├── staff-software-engineer.md orchestrator agent
│   └── system-architect.md        orchestrator agent
├── commands/                      ← slash-command entry points
│   ├── golden-path.md             /golden-path  (front door: idea → merged PR)
│   ├── product-manager/           /product-manager:{prd,prp,run}
│   ├── sse/                       /sse:{plan,dev,test,pr,run,sdd,firebase-publish}
│   ├── system-design/             /system-design:{design,review,run}
│   ├── context/                   /context:{pack,graph}
│   └── pipeline/                  /pipeline:{continue,reset}
├── shared/                        ← cross-agent guides (context-strategy.md)
├── conventions/                   ← generic conventions (overridable per repo)
├── hooks/                         ← root lifecycle hooks (session-start, prompt, postedit, postwrite, status-line, activity-pre-read)
├── scripts/                       ← root utilities (pipeline.py, activity.py, pr-monitor.py, marker.sh, preflight.sh, pack-repo.sh, graph-repo.sh)
└── runtime/                       ← generated artifacts + per-agent hooks/scripts
    ├── outputs/
    │   ├── pm/{prd,prp,tokens,.markers}/
    │   └── sse/{plan,dev,test,pr,tokens,.markers}/
    ├── hooks/
    │   ├── product-manager/       phase markers, sensor gates, eval, publish
    │   └── staff-software-engineer/
    └── scripts/
        ├── product-manager/       sensor-runner, token-phase, link-validator, confluence-publish, run-sensors.sh
        └── staff-software-engineer/  token-phase, run-sensors.sh

Read the full file on GitHub · 243 lines

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 · 243 lines · 3,723 tokens per session scan C f688b302249f

Subscribe to this mod's changes

harness-kit AGENTS.md is an instructions file published in the GitHub repository Pierry/harness-kit (3 stars, last pushed 1mo ago), licensed MIT. It adds 3,723 tokens to every session, about $0.0186 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.