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.
git clone --depth 1 https://github.com/andr-ca/agentharnessWrote 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/rules/andr-ca/agentharness/port-agent-config)<a href="https://agentmods.dev/rules/andr-ca/agentharness/port-agent-config"><img src="https://agentmods.dev/badge/rules/andr-ca/agentharness/port-agent-config/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/rules/andr-ca/agentharness/port-agent-config"><img src="https://agentmods.dev/badge/rules/andr-ca/agentharness/port-agent-config.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00112 | $0.02106 |
| Opus 5.5 | $0.00045 | $0.00842 |
| Sonnet 5.5 | $0.00022 | $0.00421 |
| Haiku 4.5 | $0.00011 | $0.00211 |
Grade A, and why
port-agent-config 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 1mo 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.
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 — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Port Agent Config
Move a project's agent instructions and skills to work with a different coding tool, without breaking what already works for the current one. Two situations, two procedures — check which one applies before doing anything.
Step 0: which situation is this?
Look for a reachable agentharness checkout with tools/generate-*.sh in it:
- This repo itself (agentharness) — the scripts are right here.
- A consumer project where
harness-link.sh initran:readlinka.claude/skills/*entry to find the harness's real path, or look for anagentharness/submodule, or a durable npm-mode copy.
Found one → Section A. Not found → Section B.
Section A — agentharness-linked: use the real generators
Don't hand-write the target file. Each generator already reuses the exact
same source (CLAUDE.md + the skill catalog), is CI-drift-checked against
its committed output, and carries the correct caveats — reconstructing its
shape by hand is strictly worse and will drift the next time CLAUDE.md
changes.
| Target tool | Generator | Produces |
|---|---|---|
| Codex CLI | tools/generate-agents-md.sh --output AGENTS.md |
AGENTS.md |
| Gemini CLI / Antigravity | tools/generate-gemini-md.sh --output GEMINI.md |
GEMINI.md |
| GitHub Copilot | tools/generate-copilot-instructions.sh --output-dir . |
.github/copilot-instructions.md + .github/instructions/*.instructions.md |
| Kilo Code | tools/generate-kilo-rules.sh --output .kilo/rules/agentharness.md |
.kilo/rules/agentharness.md |
| Cursor | tools/generate-cursor-rules.sh --output-dir . |
.cursor/rules/agentharness-router.mdc + one .mdc per skill |
| OpenCode / Zed | nothing to generate — both read AGENTS.md directly |
— |
| Claude Code | nothing to generate — CLAUDE.md is the hand-authored source |
— |
Steps:
- Run the matching command against the target project (see each script's
own
--help, ordocs/INTEGRATION.md's per-platform section when the harness checkout and the target project aren't the same directory). - Re-run the project's own content-quality check (
tools/verify-content-quality.pyhere; the consumer's equivalent elsewhere) to confirm the new file isn't silently broken. - Tell the user this is a manual step, not auto-wired into
init/update(ROADMAP.md's P1-01) — they'll need to re-run it wheneverCLAUDE.mdor the skill catalog changes. - Keep the same caveat every generated file already states: not verified against a live session of the target tool.
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.
- 1mo ago First seen · 138 lines · 112 tokens per session scan A ee0d034caa43
port-agent-config is a cursor rule published in the GitHub repository andr-ca/agentharness (1 stars, last pushed 5d ago), licensed MIT. It adds 112 tokens to every session and 2,106 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-03.
Other cursor rules, from other repositories
handoff
Export a Waybill Bundle by default, or import one when explicitly requested.
dreamd-recall
Recall lessons, decisions, and prior context from the .agent/ memory daemon. Use when starting work in a project that has a .agent/ folder, when the user references a past decision, or when you are about to make a choice that has a documented prior.
lemoncrow.execute
LemonCrow execute mode reference for Cursor.
ui-motion-workflow
Orchestrate motion-aware UI work by sequencing direction, provider choice, implementation, and browser validation.
prompt-routing
Route tasks to the correct Universal AI Engineering Prompt.
ponytail
Ponytail, lazy senior dev mode. Always pick the simplest solution that works.