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/paladini/harness-scoreWrote 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/paladini/harness-score/core)<a href="https://agentmods.dev/rules/paladini/harness-score/core"><img src="https://agentmods.dev/badge/rules/paladini/harness-score/core/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/paladini/harness-score/core"><img src="https://agentmods.dev/badge/rules/paladini/harness-score/core.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 | $0.00336 | $0.00336 |
| Opus 5.5 | $0.00134 | $0.00134 |
| Sonnet 5.5 | $0.00067 | $0.00067 |
| Haiku 4.5 | $0.00034 | $0.00034 |
Grade A, and why
core 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 18d 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.
What it actually says
- The scanner is deterministic by contract: no model calls, no network, no time-dependent output. If a proposed check needs judgment, it does not belong in the CLI — it belongs in the guide as advice.
packages/clihas zero runtime dependencies. Do not add one.- Check changes touch three files together:
packages/cli/src/score.ts,docs/guide/maturity-model.md,docs/guide/measure-and-improve.md— enforced in CI bydocs.test.tsandmaturity-sync.test.ts. - Guide edits must keep all five locales in sync — see
.cursor/rules/docs-i18n-parity.mdcanddocs.test.ts(file set + anchor parity). - Check IDs are stable public API (they are documentation anchors and appear in users' CI logs). Never renumber, reuse, or delete an ID — deprecate by making the check always-pass if truly needed.
- This repository must keep scoring L4 on
npm run scan. - External PRs: push fixes to the contributor PR head branch, not local
mainahead oforigin/main. - English on GitHub: PR/issue comments and review bodies on
paladini/harness-scoreare English only (changesets/release notes too, unless the user explicitly asks for another language). Do not post maintainer review feedback in Portuguese or other locales on public threads.
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.
- 18d ago Changed · +6 lines · +91 tokens per session 597519f98680
- 1mo ago First seen · 19 lines · 245 tokens per session scan A 7f0939d8b912
core is a cursor rule published in the GitHub repository paladini/harness-score (552 stars, last pushed 3d ago), licensed MIT. It adds 336 tokens to every session, about $0.0013 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 2026-08-30.
Other cursor rules, from other repositories
prompt-to-asset
Asset generation — route, rewrite, generate, validate. Never send transparency requests to Imagen/Gemini.
swamr-orchestrator
Swamr Orchestrator — autonomous agent swarm that builds entire projects. Invoke with @swamr-orchestrator followed by what you want built. Decomposes into tasks, selects specialist agents, runs dev↔QA loops, and delivers production-ready code.
swamr-state-manager
Swamr State Manager — tracks task progress, manages the execution pipeline, handles retries and escalation. Maintains swamr/state.json as the single source of truth for swarm progress.
read-xlsx
Reading, writing, diffing, and repairing spreadsheets (.xlsx) for AI agents via the xfa MCP server.
seed-phrase-handling
Never expose IronWallet recovery phrases or private keys.
monobehaviour-lifecycle
Correct usage of MonoBehaviour lifecycle methods.