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.
npx agentmods add instructions/enmanuelmag/agent-harness-kit/agents-mdgit clone --depth 1 https://github.com/enmanuelmag/agent-harness-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/enmanuelmag/agent-harness-kit/agents-md)<a href="https://agentmods.dev/instructions/enmanuelmag/agent-harness-kit/agents-md"><img src="https://agentmods.dev/badge/instructions/enmanuelmag/agent-harness-kit/agents-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00929 | $0.00929 |
| Opus 5 | $0.00464 | $0.00464 |
| Sonnet 5 | $0.00186 | $0.00186 |
| Haiku 4.5 | $0.00093 | $0.00093 |
Grade A, and why
agent-harness-kit AGENTS.md scanned grade A with 0 findings 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 today.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Available Research Tools
- Context7 MCP tools available: resolve library ID before querying docs
- Mintlify Index available for publisher-maintained technical documentation
- Web search available for current information outside documentation indexes
AGENTS.md — @cardor/agent-harness-kit
Read this file first. It is the navigation map for every AI agent working in this repository.
Project
@cardor/agent-harness-kit — A CLI and MCP tools for LLM providers
Health check (run before making codebase changes)
bash health.sh
If it exits non-zero, stop and report the issue. Do not proceed with codebase changes until health is green.
Harness data (source of truth)
| File | Purpose |
|---|---|
.harness/harness.db |
SQLite: all tasks, actions, file changes, tool calls |
.harness/current.md |
Markdown fallback — read this if MCP server is unavailable |
.harness/feature_list.json |
Human-editable task seed list |
MCP tools (preferred)
The harness exposes tools via MCP server on port 3742. Use these instead of reading files directly.
actions.start taskId agent → start an action, returns a numeric actionId
actions.write actionId section text → record a section (result, blockers, ...)
actions.record_tool actionId calls[] → batch-log tool calls to the Tools dashboard (array, min 1)
actions.record_file actionId files[] → batch-log file touches to the Files dashboard (array, min 1)
actions.complete actionId summary → close the action
actions.list taskId [agent] [status] [limit] [cursor] → compact newest-first action index with pagination
actions.get_by_id actionId → single action with section index (no content)
actions.sections_list actionId [types] [limit] [cursor] → compact section index with type filtering
actions.sections_get sectionId [length] [offset] → ranged section content reader
actions.handoff.get taskId [recipient] → newest completed canonical handoff
actions.handoff.write actionId recipient ... → recipient-directed bounded handoff
tasks.add title [slug] [description] [acceptance] → create a new task from natural language
tasks.get [status] → list tasks (pending | in_progress | done | blocked)
tasks.claim id → atomically claim a pending task
tasks.update id status → change task status
tasks.acceptance.update criterionId → mark an acceptance criterion as met
docs.search query → search ./docs for relevant content
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.
- today Changed · +11 lines · +149 tokens per session 83be7672ced7
- 4d ago First seen · 81 lines · 780 tokens per session scan A 021c49ad2695
agent-harness-kit AGENTS.md is an instructions file published in the GitHub repository enmanuelmag/agent-harness-kit (180 stars, last pushed today), licensed Apache-2.0. It adds 929 tokens to every session, about $0.0046 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
agent-lab-opencode AGENTS.md
AGENTS.md instructions for SkS-Other/agent-lab-opencode, covering grace framework - project engineering protocol, keywords, annotation, core principles and 1. never write code without a contract.
hoop CLAUDE.md
Claude Code instructions for hoophq/hoop, covering claude.md, project overview, toolchain & prerequisites, architecture and module breakdown.
gini-agent AGENTS.md
AGENTS.md instructions for Open-Curiosity/gini-agent, covering gini agent instructions, shape, adrs, boundaries and branches.
rome CLAUDE.md
Claude Code instructions for rome-os/rome, covering rome, playbook and traps.
zapcode CLAUDE.md
Instructions for TheUncharted/zapcode, covering claude.md, git, read agents.md first, codebase orientation and how to add a new language feature.
ralph CLAUDE.md
Claude Code instructions for SantanderAI/ralph, covering ralph — instrucciones del proyecto and paridad bash ↔ powershell (importante).